1What group efficiency isThe group against the sum of its piles.
Piles in a group stand close enough for the zones of soil each of them stresses to overlap. A friction pile in a clay then carries less in the group than it would alone, and the group efficiency η says by how much: it is the ultimate capacity of the group over the sum of the ultimate capacities of its piles acting alone,
n is the number of piles and Qu the capacity of one pile alone. The equations on this page are empirical: they take the number of piles and the ratio of their width D to their centre-to-centre spacing s, and give η as one number. They were written for friction piles in clay; piles in sand follow the rules further down, and a group in soft clay at close spacing must also be checked as one block (see pile group block failure).
2The Converse–Labarre equationNeighbouring pairs, each costing an angle.
The Converse–Labarre equation (Bolin, 1941) counts the neighbouring pairs of piles. A group of n1 rows with n2 piles in each row has (n1 − 1)n2 pairs in one direction and (n2 − 1)n1 in the other, and each pair costs the angle θ, in degrees, over 90:
When the spacing differs in the two directions, each set of pairs takes the angle of its own spacing, which is how the GeoTerva pile group module writes it:
η depends on s/D alone, not on the size of the piles: a 3 × 3 group at s = 3D has η = 0.727 whatever their diameter. For a triangular (staggered) arrangement the module counts the neighbouring pairs of that layout in the same way; every pile then has more neighbours, and η is lower than that of a rectangular grid at the same spacing.
3The Los Angeles and Seiler–Keeney equationsTwo other forms, for comparison.
Two other equations are often quoted beside it. The Los Angeles group-action equation counts the diagonal pairs as well:
The equation of Seiler and Keeney (1944) was written with the spacing s in feet, so it depends on the spacing itself and not only on s/D:
It has a pole at s = 1 ft and gives meaningless values at small spacings; GeoTerva does not give it below 2 ft (0.61 m). For a rectangular group the module prints all three equations and uses Converse–Labarre for the clay layers.
4Tables of ηConverse–Labarre for the usual layouts, and the three equations side by side.
| Group | s/D = 2 | s/D = 2.5 | s/D = 3 | s/D = 3.5 | s/D = 4 | s/D = 5 | s/D = 6 |
|---|---|---|---|---|---|---|---|
| 2 × 2 | 0.705 | 0.758 | 0.795 | 0.823 | 0.844 | 0.874 | 0.895 |
| 2 × 3 | 0.656 | 0.717 | 0.761 | 0.793 | 0.818 | 0.853 | 0.877 |
| 2 × 4 | 0.631 | 0.697 | 0.744 | 0.779 | 0.805 | 0.843 | 0.869 |
| 3 × 3 | 0.606 | 0.677 | 0.727 | 0.764 | 0.792 | 0.832 | 0.860 |
| 3 × 4 | 0.582 | 0.657 | 0.710 | 0.749 | 0.779 | 0.822 | 0.851 |
| 3 × 5 | 0.567 | 0.645 | 0.700 | 0.740 | 0.771 | 0.816 | 0.846 |
| 4 × 4 | 0.557 | 0.637 | 0.693 | 0.734 | 0.766 | 0.812 | 0.842 |
| 4 × 5 | 0.542 | 0.625 | 0.683 | 0.725 | 0.758 | 0.805 | 0.837 |
| 5 × 5 | 0.528 | 0.612 | 0.672 | 0.717 | 0.750 | 0.799 | 0.832 |
| 6 × 6 | 0.508 | 0.596 | 0.659 | 0.705 | 0.740 | 0.791 | 0.825 |
η depends only on the layout and on s/D. Values computed by GeoTerva with the function its pile group module uses.
| s/D | s (m) | s (ft) | Converse–Labarre | Los Angeles | Seiler–Keeney |
|---|---|---|---|---|---|
| 2.0 | 1.00 | 3.28 | 0.582 | 0.662 | 0.603 |
| 2.5 | 1.25 | 4.10 | 0.657 | 0.730 | 0.703 |
| 3.0 | 1.50 | 4.92 | 0.710 | 0.775 | 0.765 |
| 3.5 | 1.75 | 5.74 | 0.749 | 0.807 | 0.808 |
| 4.0 | 2.00 | 6.56 | 0.779 | 0.831 | 0.839 |
| 5.0 | 2.50 | 8.20 | 0.822 | 0.865 | 0.881 |
| 6.0 | 3.00 | 9.84 | 0.851 | 0.887 | 0.908 |
Seiler–Keeney takes s in feet and is not given below 2 ft. Values computed by GeoTerva.
5Piles in sandDriven piles compact the sand; bored piles do not.
In a sand the equations above do not apply. Driving piles into a loose or medium sand compacts it, and in the group tests of Kishida and Meyerhof (1965) a group of driven piles carried as much as the sum of its piles, or more; only in a dense sand, with φ′ above about 40°, did a close spacing reduce the capacity. GeoTerva takes η = 1 for driven and jacked piles below φ′ = 40°, the curve of those tests for φ′ = 45° and above, a linear blend between the two, and never a value above 1. A bored pile displaces no soil; for shafts in a granular soil AASHTO LRFD takes
with 0.65 at closer spacings and 1.0 beyond 4D.
| s/D | Driven, φ′ = 38° | Driven, φ′ = 42.5° | Driven, φ′ ≥ 45° | Bored (AASHTO LRFD) |
|---|---|---|---|---|
| 1.5 | 1.000 | 0.809 | 0.619 | 0.650 |
| 2.0 | 1.000 | 0.847 | 0.695 | 0.650 |
| 2.5 | 1.000 | 0.879 | 0.758 | 0.650 |
| 3.0 | 1.000 | 0.906 | 0.813 | 0.767 |
| 3.5 | 1.000 | 0.929 | 0.859 | 0.883 |
| 4.0 | 1.000 | 0.949 | 0.898 | 1.000 |
| 5.0 | 1.000 | 0.978 | 0.957 | 1.000 |
| 6.0 | 1.000 | 0.996 | 0.991 | 1.000 |
| 7.0 | 1.000 | 1.000 | 1.000 | 1.000 |
Driven: after the group tests of Kishida and Meyerhof (1965), held at 1 at most; bored: AASHTO LRFD. Values computed by GeoTerva.
6Worked exampleTwelve driven piles in clay.
Twelve driven concrete piles, D = 0.50 m and L = 18 m, stand in 3 rows of 4 at s = 1.50 m (s/D = 3.0) in a clay with cu = 60 kPa.
- One pile alone (α method, α = 0.62 for cu/pa = 0.60; base 9cuAb): Qs = 0.62 × 60 × π × 0.50 × 18 = 1051.8 kN, Qp = 9 × 60 × 0.1963 = 106.0 kN, Qu = 1157.8 kN. The 12 piles alone add up to ΣQu = 13894 kN.
- Converse–Labarre: θ = tan−1(0.50/1.50) = 18.43°, and the group has (3 − 1) × 4 + (4 − 1) × 3 = 17 neighbouring pairs: η = 1 − 18.43 × 17 / (90 × 3 × 4) = 0.710, so Qg(u) = 0.710 × 13894 = 9862 kN.
- Los Angeles: η = 0.775, Qg(u) = 10763 kN. Seiler–Keeney (s = 4.92 ft): η = 0.765, Qg(u) = 10633 kN.
The three equations differ by about 6 % of the sum of the piles; Converse–Labarre, the lowest of the three here, is the one the module uses for a clay. The same twelve piles in a triangular arrangement, every pile 1.50 m from its neighbours, have η = 0.607. As one block of 5.00 m × 3.50 m × 18 m the group would carry 27522 kN, far more than the piles: at this spacing block failure does not govern (see pile group block failure). How the load of the cap is shared among the twelve piles is the subject of the load on each pile.
7QuestionsShort answers.
Which equation does the GeoTerva pile group module use?
Converse–Labarre for every clay layer and for a pile base in clay; the rules for driven or bored piles for a sand layer. The Los Angeles and Seiler–Keeney values are printed beside it for comparison. In layered ground each layer takes the efficiency of its own soil, so the efficiency of the group is a weighted mean over the layers.
Does the efficiency depend on the length of the piles?
Not in these equations: they take the number of piles and s/D (Seiler–Keeney the spacing in feet). The length of the piles enters the block check, which compares the group as one block with the piles.
Can a pile group efficiency be greater than 1?
Driven piles in a loose or medium sand compact the soil, and group tests have shown efficiencies above 1. The gain depends on the initial density and on the order of driving; GeoTerva does not count on it and holds η at 1 at most.