Negative skin friction and downdrag: when soil loads the pile

Negative skin friction develops when the soil around a pile settles more than the pile itself: the side shear reverses direction and drags the pile down instead of holding it up. The accumulated force — the dragload — peaks at the neutral plane, where soil and pile settle equally. It adds to the structural load in the pile; it does not reduce the geotechnical bearing capacity.

The mechanism: shear direction follows relative movement

Skin friction has no fixed sign. The shear stress on a pile shaft acts in whatever direction the soil moves relative to the pile. In the ordinary case the pile moves down relative to the soil under applied load, so the soil resists — positive skin friction, the familiar half of skin friction vs end bearing. But when the surrounding soil is consolidating, the soil moves down relative to the pile over the upper part of the shaft. The same interface, the same unit friction — now pointing downward, loading the pile rather than supporting it.

Two facts make this hard to avoid. First, it takes only a few millimetres of relative movement to fully mobilize shaft friction, so even "small" long-term settlement of a consolidating layer is enough to reverse the shear over years of service. Second, the magnitude of the negative unit friction is essentially the same as the positive unit friction you would have counted on in that layer — so a thick consolidating deposit can hang a substantial force on the pile.

When downdrag happens

• New fill over soft clay. The classic case: a few metres of embankment or site-grading fill placed over a compressible deposit starts consolidation that continues for years after the piles are driven.
• Groundwater lowering. Dewatering or long-term drawdown increases effective stress throughout the profile and re-starts consolidation — no new surface load required.
• Liquefaction reconsolidation. After an earthquake, liquefied sand layers reconsolidate as excess pore pressures dissipate, settling around any pile that passes through them.
• Under-consolidated or organic ground. Recent fills, hydraulic fills, and organic soils that are still settling under their own weight load every pile installed through them.

The neutral plane

Somewhere down the shaft, the settlement of the soil equals the settlement of the pile. That depth is the neutral plane. Above it, soil settles more than the pile — skin friction is negative and the axial force in the pile grows with depth. Below it, the pile settles more than the soil — friction is positive and the axial force decays toward the toe. The axial force diagram therefore has its maximum at the neutral plane:

P_max (at neutral plane) = P_dead + Q_neg

and in equilibrium:  P_dead + Q_neg = R_side,positive (below NP) + R_toe,mobilized

The neutral plane is also a force balance: it sits at the depth where dead load plus accumulated dragload equals the positive resistance still available below. A stiff toe-bearing pile pushes the neutral plane deep (often near the bearing layer) and collects a large dragload; a compressible friction pile finds its neutral plane higher with less. The same relative-movement logic drives the pile's long-term settlement: the pile settles with the soil at the neutral plane, which is why downdrag problems are usually settlement problems.

The design check — and the classic misconception

The mistake that appears in calculations everywhere: subtracting dragload from the pile's allowable capacity, i.e. Q_allow − Q_neg ≥ P_design. Per Fellenius's neutral-plane (unified) method, that double-counts against you and misses the real checks. Dragload cannot reduce geotechnical capacity, for a simple kinematic reason: at plunging failure the pile moves down relative to the soil everywhere along the shaft, so all skin friction is positive at the moment capacity is mobilized — the negative friction has vanished from the failure mechanism. The checks are:

• Geotechnical: dead plus live load against the full axial capacity (all friction positive plus toe), with dragload excluded.
• Structural: dead load plus dragload against the pile section's structural strength at the neutral plane, where the axial force peaks. This is where downdrag genuinely bites, especially for long piles to a hard bearing layer.
• Settlement: the pile settles with the soil at the neutral plane; check that this — plus the pile's own compression — is tolerable.

A related subtlety: transient live load and dragload are not additive. A short-term live load compresses the pile and momentarily reverses the relative movement in the negative zone, shedding dragload as it arrives — so combining full live load with full dragload in one load case double-counts.

Mitigation

• Bitumen coating. A coat over the negative-friction zone lets the soil slide by at a small fraction of the uncoated shear. Effective, but fragile during driving and dependent on temperature and coating thickness.
• Sleeves or double casings. Isolate the shaft from the settling soil mechanically over the affected depth.
• Preloading / surcharging. Get the consolidation over with before the piles are installed — often the cheapest fix when schedule allows.
• Sequence the work. Place fill early and drive piles after the bulk of consolidation is complete, rather than the reverse.
• Design for it. Frequently the right answer is simply a pile section with enough structural capacity at the neutral plane — dragload is a load, and loads can be carried.

How PileCalc treats a downdrag zone

PileCalc's axial engine handles downdrag at two levels, and it is worth being precise about what each one does.

In the capacity calculation (NAVFAC DM-7.02 method), you mark depth intervals as negative zones. Segments inside a negative zone contribute nothing to downward or uplift side resistance; instead the engine accumulates a dragload from the same unit friction it would otherwise have credited:

Q_neg = Σ K_neg · f · perimeter · Δz    (over the negative zone)

Here f is the NAVFAC unit side friction for the layer (β-type K·σ′ᵥ·tan δ in cohesionless soil, α-type adhesion in clay) and K_neg is a 0–1 reduction factor — set it below 1 to represent, say, a bitumen coating. The dragload is reported as a separate result alongside the capacity, with its per-segment distribution, and is not subtracted from the geotechnical capacity — consistent with the neutral-plane reasoning above. Zero-resistance zones (soil you trust for neither support nor drag, such as a pre-drilled interval) are available separately: they remove side friction without adding dragload. Note that in this mode the negative zone is your input — the engine does not locate the neutral plane for you.

For the full mechanics, the t-z load-transfer solver accepts an imposed free-field soil-settlement profile u_soil(z) and solves the axial bar on nonlinear side springs,

d/dz(EA · du/dz) + perimeter · τ = 0,    τ driven by (u_soil − u_pile)

with the side shear at every depth mobilized by the relative movement between soil and pile through the t-z curve, and end bearing as a nonlinear spring at the tip. Feed it a consolidation settlement profile and the neutral plane emerges from the solution: the shear reverses sign where relative movement changes direction, and the computed axial force distribution peaks there. That gives you the maximum force for the structural check and the pile settlement for the serviceability check, from one analysis. The normalized t-z and q-w curves it runs on are the same family behind the load–settlement output in the pile settlement calculator.

Common questions

Does negative skin friction reduce pile capacity?

No. Dragload is a structural load, not a capacity deduction. At geotechnical failure the pile moves down relative to the soil along its whole length, so all skin friction acts upward at that moment and the dragload disappears from the failure mechanism. Downdrag is checked against the pile's structural strength at the neutral plane and against settlement — not against bearing capacity.

Where is the neutral plane on a pile?

At the depth where soil settlement equals pile settlement — equivalently, where dead load plus accumulated dragload balances the positive resistance available below. Stiff piles bearing on a hard layer have deep neutral planes and large dragloads; compressible friction piles have shallower neutral planes. Its location shifts with the loading, since more dead load pushes the balance point upward.

Should live load be added to dragload?

No. Transient live load and dragload do not act at full value simultaneously. When a short-term load arrives, the pile compresses and moves down relative to the soil in the upper zone, momentarily converting negative friction back to positive. Combining full live load with full dragload in one case double-counts; the structural check uses sustained load plus dragload.

How much does bitumen coating reduce negative skin friction?

A properly applied bitumen coat lets consolidating soil creep past the shaft at a small fraction of the uncoated interface shear — commonly modeled as a reduction factor on the unit friction in the coated zone (K_neg in PileCalc). Field effectiveness depends on coating thickness, temperature, and surviving the driving process, so specify and inspect it carefully.

If you want to see the numbers on a real profile, run your pile in the free pile capacity calculator, then open the full app to add negative zones and read the dragload and its distribution directly from the segment-by-segment output.