Block failure in clay: the check Converse-Labarre misses

Geometric Converse-Labarre is screening, not clay truth. For a tight friction-pile group in cohesive soil, an efficiency formula alone can overestimate capacity — you also need the equivalent-block check. Theory and the published η table live on the pile group efficiency hub. This page isolates the block mode that Converse-Labarre does not compute. Not a PE stamp, not a design example for a real project.

What Converse-Labarre misses

Converse-Labarre is layout-only. It returns a reduction factor η from rows, columns, diameter, and spacing. It does not know undrained strength, embedment, or whether the group can fail as one prism of soil and piles. The hub already says that for closely spaced groups in cohesive soil, an efficiency factor is not enough.

Geometric screening at s/d = 3 for a 3×3 (hub worked example: 0.6 m piles at 1.8 m centres) gives η ≈ 0.73. That is nine piles worth about 6.5–6.6 singles on the geometric path — screening only, not the clay block result. Work the formula on Converse-Labarre equation or the group efficiency lab; then remember the block check is separate.

The equivalent block

Check failure of the group as an equivalent block: a prism of soil and piles with plan dimensions B_x × B_y (outside face to outside face) and the piles' embedded length L. Hub capacity is perimeter shear plus base bearing:

Q_block = 2 · L · (Bₓ + Bᵧ) · c̄ᵤ  +  N_c · cᵤ · Bₓ · Bᵧ      (N_c ≤ 9)

Two details are easy to get wrong and are already on the hub. The perimeter shears soil on soil, not pile on soil — so the full undrained strength c̄ᵤ applies along the sides (no Tomlinson adhesion reduction α), and in cohesionless layers the full friction angle φ applies rather than a reduced interface angle δ. Design capacity is the lesser of the two modes: min(n · Q_single, Q_block).

When the block takes over

At wide spacing the block is huge and the individual piles govern. Squeeze the spacing and the block takes over — that crossover is what group efficiency is really measuring in clay. Below about 2.5d, expect both a heavy geometric penalty and a governing block in cohesive soil. Space at ≥ 3 diameters when you can.

This post does not invent a numerical Q_block for a site. The hub publishes the formula and the qualitative crossover; without a published L, c̄ᵤ, and plan size for a named case, treat the check as a required mode comparison, not a worked capacity for a real project.

What PileCalc releases today

PileCalc's released vertical-group tool is the driven-sand path (η = 1; USACE EM 1110-2-2906). The block is printed as a diagnostic. Clay, rock, mixed, and weak-layer block-governing cases fail closed. Details and caveats stay on the hub and in the pile groups docs.

Students and engineers should still run Converse-Labarre in the free pile group efficiency calculator, then treat the clay block check as a separate physical mode — not a substitute for engineering judgment.

Keep reading

• Live calculator — Converse-Labarre and the block comparison in the browser
• Pile group efficiency hub — formulas, η table, block failure, lateral p-multipliers
• Converse-Labarre equation — worked reduction factors from the hub table
• Lab: pile group efficiency — hand drill for the hub 3×3 s/d = 3 case

Sources: FHWA Soils and Foundations Reference Manual Vol. II (axial group efficiency and equivalent-block); WSDOT WA-RD 827.1 Eq. 2.29 (Converse-Labarre expression used on the hub). Engineering judgment stays with the user.