Load transfer on piles, construction tolerances, and support stiffnesses

PUBLI

The relationship between the structural works package and the special foundations package relies largely on the load transfer schedule which, beyond a simple listing, can play a structuring role in the project’s design and in managing the interface between the packages.


When it specifies the allocation of pile execution tolerances and the mechanical assumptions at the interface, the load transfer schedule helps to clarify the boundary between the packages and to secure the design.
Through a topic that is sometimes not explicitly addressed, this article highlights real geometric, structural, and contractual impacts that are worth documenting from the outset.

 

 

Pile execution tolerances

 

In France, the execution standards for piles allow by default relatively favorable construction tolerances for special foundation contractors:

  • NF EN 1536, which concerns bored piles, provides:
    • 10 cm in plan position
    • 2 cm/m in inclination
  • NF EN 12699, which concerns screw piles, provides:
    • 10 cm in plan position
    • 4 cm/m in inclination
  • The informative NF P94-262 proposes by default, when the technology is not fixed:
    • 15 cm in plan position
    • 3 cm/m in inclination

In addition, these tolerances apply from the working platform used by the piling machine. Consequently, the inclination deviation may amplify the plan deviation when piles are constructed before final excavation, from a higher platform.

 

 

elevated or low plaform of piling

 

This high-platform configuration may occur during construction of retaining walls (secant, tangent, contiguous piles, soldier pile walls…), or more generally if access, sequencing or schedule constraints do not allow the piling rig to descend to the final excavation level.

 

 

Structural drawings from the design office

 

Plan position and inclination deviations are part of site reality; accounting for them early contributes to robust design.

From a geometric standpoint, identifying fixed dimensions in the drawings and providing the architect with the correct clearances and dimensional allowances ensures the compliance of the structure, provided that the piles themselves comply with tolerances.

Structural drawings produced using nominal dimensions, without considering the effects of tolerances, may lead to problematic situations during construction: piles outside property limits, non‑compliant internal dimensions (parking space depth, drive aisle width…), etc. Anticipating them reduces this risk.

These deviations may also be worsened by service‑stage deformations of special foundations, sometimes reaching several centimeters (pile head displacement, wall head or mid‑height deformation…).

In particular, the elevation—discussed previously—of the future working platform for piling contractors is one of the fundamental assumptions that the structural design office benefits from knowing and drawing to ensure proper design and documentation.

 

 

Special tender documents

 

Tightening execution tolerances

 

Special contractual clauses for the piling package may, when necessary, specify tighter normative tolerances than those previously mentioned, within reasonable limits (NF P94‑262 R.1(2) and R.3(5)).

 

“Assigning responsibility” to the piling contractor

 

The contract may also require that the piles absorb the effects resulting from their plan and inclination tolerances.

The structural design office then translates this design choice into the ultimate limit state load path provided to the special foundations design office.

Note: geometric imperfection effects are generally ignored at serviceability limit state, in accordance with P94‑262 R.1.(5) and, more broadly, EC2 §5.2 2(P) and §5.2(3).

For several reasons, it is often simpler to assign to the piles the effects of their own tolerances:

  • Piles are designed in an axisymmetric manner and can more easily accommodate tolerances in all directions.
  • Piles are often sized at SLS for diameter, providing “reserve” capacity at ULS.
  • The effect of bending moments and horizontal forces at pile head remains localized and dissipates within a few meters.

Meanwhile, supported structures are usually designed at ULS, are rarely axisymmetric, and incorporating pile tolerances in all possible directions is often globally more expensive.

 

engineering team choice for tolerance of piles

 

In the load path, instead of directly pre‑calculating imperfection effects, a relation‑based formulation (e.g., M = 0.1N…) or a generic statement clarifies the assumptions and prevents the special foundations office from combining non‑physical loads (e.g., {Nmin and M = 0.1Nmax}), or alternatively forgetting certain directions of imperfections.

This design choice—assigning tolerance effects to piles—also benefits from being explicitly stated in the written documents for the structural and foundation packages to prevent this assumption from being “lost” during pricing or during the production of the execution load path.

 

 

Default interface between structural works and special foundations

 

The previous section mentions “assigning responsibility” to the piling package.

Indeed, in the absence of any clarification in the load path or written documents, the structural design office places on the structural works package the responsibility for the effects of position and inclination errors introduced by the piling contractor, in all directions and magnitudes (within tolerances).

On site, the piling contractor is never responsible for the structural design of the entire project and relies solely on the provided load path and execution tolerances.

When schedule allows, assigning positional errors to the supported structure might permit “optimization” of the reinforcement of beams and slabs based on as‑built pile locations, but such opportunities are rare.

In addition, as‑built pile location drawings are challenging to obtain, and measuring pile inclination is even more difficult.

 

Variants on pile diameters and platform levels

 

When reviewing tender documents for a bid submission, special foundation contractors analyze the optimal technical‑economic solution based on experience, expertise and proprietary methods.

They often seek to optimize the number and diameter of piles relative to the preliminary sizing established in the geotechnical G2 PRO mission, relying on specific techniques allowing higher concrete working stresses than standards.

They may also propose different platform elevations than those indicated in the design.

Finally, depending on their methods, they may also commit to lower tolerances than those specified.

Depending on the previously discussed topics, these adjustments may have significant impacts on other packages; aligning them early with the structural engineer facilitates offer analysis and overall coherence.

 

 

Structural design on piles

 

Consistent design choices

 

As mentioned previously, it is the structural design office that chooses which elements must absorb the effects of pile plan and inclination errors, including:

  • implicitly assigning them to the structural works package as long as they remain within tolerances
  • explicitly assigning them to the piles if desired

More subtly, Eurocode 2 encourages a structural analysis that reflects, as far as possible, the most probable behaviour of the structure. Consistent choices with the relative stiffness between pile and structure help reflect realistic behaviour:

  • example 1: if a pile supports a wall beam, the beam is naturally the element that will absorb the in‑plane positional deviation
  • example 2: if the supported structure is much less stiff than the pile, then the pile should be specified to absorb tolerance effects
  • example 3: if a pile supports a continuous beam stiffer in bending than the pile, the beam provides partial fixity that distributes the deviation based on relative stiffness

 

 

The load path may specify by families the boundary conditions (pinned, partially fixed, fully fixed), the relevant stiffness parameters and the positional and inclination errors to integrate in each direction.

It remains possible to adopt conservative approaches, easier to express but inevitably leading to oversizing—either piles or supported structures.

Regardless, a lack of specification must be a deliberate choice. Structural drawings must then show a system actually capable of absorbing pile deviations (which is not the case in example 2), and the written documents must explicitly state this assumption.

 

Pile stiffness vs structural stiffness

 

We have just mentioned the possibility of accounting for a flexural stiffness at the pile head, exerted by the superstructure when it resists rotation of the pile head in the presence of an applied moment (or an eccentricity).

Considering this partial stiffness is sometimes useful or necessary to optimize the design or to reduce displacements.

This concept essentially reflects the idea of a perfect fixity between the pile head and the supported structure.


Such perfect fixity is necessarily reciprocal and is generally represented in each engineering office’s model by a partial fixity.

  • in model 1 (structural analysis), the pile head is modelled with a partial rotational stiffness kϑ,P applied by the pile to the structure
  • in model 2 (pile structural analysis), the pile head is modelled free in translation and partially restrained in rotation with stiffness kϑ,S applied by the supported structure

This configuration is illustrated below in 2D:

 

relationships between partial rotational stiffness at pile head and structure

 

Important note: kϑ,P ≠ kϑ,S

Let kϑ,G1 and kϑ,G2 be the stiffness of beams on left and right, kϑ,C that of the column, Mtot the total moment, and ϑ the rotation.

Under perfect fixity, all elements share the same rotation ϑ at the node.

The stiffness applied by the supported structure is thus kϑ,S = kϑ,G1 + kϑ,G2 + kϑ,C.

The four stiffnesses act in parallel, so Mtot = kϑ,S·ϑ + kϑ,P·ϑ.

The pile reactions in model 1 are Ry, Rz and M1 = kϑ,P·ϑ.

Model 2 is loaded by T2, N2 and M2:

  • N2 = –Rz
  • T2 = –Ry
  • but M2 ≠ –M1

Because model 2 includes structural stiffness, the applied moment is Mtot, not –M1. Using ϑ = M1/kϑ,P, we get:

  • M2 = – (1 + kϑ,S/kϑ,P) · M1

 

Note: Modeling the pile using independent stiffnesses (in particular horizontal stiffness and rotational stiffness) is a common first approach, and it is easy to implement in a finite element model.
Strictly speaking, however, the behavior of a pile subjected to lateral loads cannot be represented by uncoupled stiffnesses: the force–displacement relationship should be described by a non‑diagonal stiffness matrix, reflecting the intrinsic coupling between horizontal translation and rotation at the pile head.

 

 

Why not simply apply –M1 at the pile head and ignore structural rotational stiffness?

 

The structural model (model 1) is a macroscopic elastic, linear and idealized representation, distributing forces and reactions.

The pile model (model 2) includes geometric imperfections (e·N2, i·N2), non‑linear effects in concrete (cracking, plasticity), soil yielding and second‑order effects.

Under such effects, kϑ,S has a significant influence on the moment and shear envelopes along the pile.

kϑ,S also plays a key role in evaluating the horizontal stiffness ky,P contributed by the pile.

Practically, including kϑ,S simplifies the treatment of geometric imperfections: model 2 naturally distributes the right share of effects, whether in cases 1, 2 or 3 above.

Strictly, the loop closes when the moment resisted at pile head (kϑ,S·ϑ) from model 2 is returned to the structural design office, which includes it in the sizing of beams/columns and draws the actual reinforcement allowing the perfect fixity between elements.

 

Effect on piles of column and wall inclination tolerances

 

Column and wall execution tolerances are much smaller than pile tolerances and less impactful, but still worth discussing.

NF EN 13670 CN §10.4 specifies tolerances depending on geometry, typically: 0.5 cm/m inclination and 1.5 cm plan deviation.

We focus here on the effect of column inclination error, illustrated below:

 

mechanical diagrams of braced and unbraced RC columns

 

Two cases arise:

 

example 1: braced column

 

Here, the column is “braced” in EC2 terms: the upper slab is horizontally restrained by a shear wall aligned with the force direction.

The inclination creates a horizontal force +H in the slab acting as a diaphragm, which re‑aligns the axial load N. The opposite force –H appears at the lower slab, ultimately transmitting a vertical load into the pile.

In principle, –H reaches the pile. However, diaphragm and wall action transmit +H downward, cancelling it with –H.

Although diaphragm rigidity is idealized, horizontal forces from inclined braced columns can generally be ignored.

 

example 2: unbraced column

 

Here, no shear wall exists downstream. The column must be laterally stable itself; EC2 calls it “unbraced”.

The inclination does not produce a horizontal force but a bending moment accompanying load N down to the pile. At pile head, the inclination represents an eccentricity 0.005·h, added to the 1.5 cm positional error.

For h = 3 m, this gives 3 cm — a non‑negligible value compared with the 10 cm pile tolerance, and may be amplified by second‑order effects.

 

example 3: multiple parallel unbraced columns

 

An unbraced column rarely behaves alone. Assuming all columns lean in the same direction is rarely realistic.

The diaphragm relieves the most inclined columns by redistributing loads to the straighter ones.

EC2 §5.2(5) allows reducing the effective inclination using coefficient αm, which equals 0.87 for 2 columns and tends toward 0.7 for many columns. Another approach is to keep maximum inclination and introduce a stabilizing force R representing diaphragm contribution.

 

illustration of multiple parallel unbraced columns

 

Effect on piles… of pile inclination tolerances

 

The previous discussion on columns helps introduce the forces induced by pile inclination tolerances on the piles themselves.

Unlike columns, actual pile inclinations are “invisible” and difficult to measure; the only way to account for them is to integrate tolerances proactively in the calculation.

These tolerances are typically 4 to 6 times larger than column tolerances: their effects are significant and may amplify other defects and contribute to second‑order effects.

Moreover, contrary to intuition, a pile restrained by a slab is not braced: piles often provide the only means of transmitting horizontal actions to the ground.

In rare cases of slabs on grade with piles, soil–slab friction may help butt pile heads, but generally remains insufficient.

Although pile heads are unbraced, slab diaphragm effects may again reduce the effective pile inclination by up to 30% using αm.

The following diagram shows a building with asymmetric earth pressure, where ϑ0 is the specified tolerance and ϑi the reduced inclination used in pile design.

 

example of building with asymmetric soil pressure on piles

 

 

This section illustrates the limits of “implicit” interface rules between structural works and special foundations:

If the structural design office does not introduce, in the ULS load path, a horizontal action of type H = αm·ϑ0·N for each pile, then the foundations engineer will assume the structural works package must absorb these effects.

But this is often unrealistic, since horizontal forces can only be transmitted to the ground by the piles.

 

Summary through an example

 

We summarise the above using an example of a suspended slab supported by beams and piles, studied using two strategies.

The example and its handling under the two strategies are deliberately idealized.

 

 

Strategy 1

 

strategy 1 with bidirectional beams absorbing pile tolerances

 

Pile heads are modelled pinned. Their geometry, connections and reinforcement are not designed to be fixed to the supported structure.

Positional tolerances are handled by the supported structure, consistent with default interface rules. The design office explicitly states this and documents the assumptions to avoid ambiguity.

In this strategy, a bidirectional beam network makes this assumption feasible: beams absorb pile positional errors in all directions and within tolerances.

Inclination tolerances, however, are assigned to the piles, as the opposite assumption is unrealistic for the project.

 

Load path provided by structural design office and attached to contract documents

 

load path table for strategy 1

 

Specifications included in both structural and foundation contract documents

 

Actions from positional tolerances within limits are included in supported‑structure design for the structural package, in all directions.

Actions from inclination tolerances are included in pile design for the foundations package.

Tolerances at tender stage: ϑ0 = 2 cm/m, e0 = 10 cm. Piles are assumed constructed after excavation (low platform). Translational stiffness assumptions are included in the load path.

Any modification must be explicitly stated in the offer.

 

Strategy 2

 

strategy 2: exploiting structural stiffness and partial fixity

 

The slab is supported on two sides by beams acting as pile caps. Beam, wall and pile stiffnesses are included in both analysis models.

The load path directly provides stiffness assumptions and required imperfections. They are transcribed in the contract documents.

 

Load path provided to the foundations engineer

 

load path for strategy 2

 

Specifications included in both packages

 

Positional tolerance effects are included:

  • either in pile design,
  • or in supported‑structure design

The load path specifies case‑by‑case assumptions.

Inclination tolerance effects are included in pile design for the foundations package.

Tolerances considered at tender stage: ϑ0 = 2 cm/m, e0 = 10 cm. Piles are assumed built after excavation (low platform). Rotational and translational stiffness assumptions are documented.

Any modification must be explicitly stated in the offer.

 

 

Conclusion

 

Execution tolerances for piles, deviations in inclination or positioning, and platform levels are not minor details: they directly influence the geometry of a structure, the design logic, the sizing of structural elements, and the distribution of responsibilities between the Structural Works package and the Special Foundations package.
A load transfer schedule is therefore not just a simple exported table of values, but a true design document that defines the geometric and mechanical assumptions on which the entire interface between the two packages will be built. Beneficially, the load transfer schedule can also clarify the assumed flexural stiffness interactions between the structure and each pile.

By stating the selected tolerances, the boundary conditions at the pile head, and the associated effects (H, eccentricities, relative stiffnesses) within the load transfer schedule and the project specifications, the structural engineer secures the project, prevents interpretation inconsistencies, and ensures that each stakeholder operates within a clear and controlled framework.

Conversely, the absence of such detail may lead to an implicit and less suitable distribution of effects. In a context where contractor alternatives, specific construction methods, and techno‑economic optimizations are common practice, the quality of the load transfer schedule becomes a key factor in preserving structural coherence, contractual balance, and the overall robustness of the structure.

 

 

 

Share this article

React, exchange with the author, view the discussions