Question 4 of 9: When to prefer the λ method over the α or β method

Question 4: When to prefer the λ method over the α or β method (7 marks)

Question text not reproduced: the examination questions are © Engineers and Geoscientists BC. Open the official past paper (linked at the top of this page) to read the question, then follow the worked solution below.

All three methods estimate the same quantity — the unit skin friction on a pile shaft in clay — but they reach it from different starting points, and the choice among them is a choice about which of the available data can be trusted.

What each method assumes. The α method is a total-stress method: f = α c_u, where the adhesion factor α falls from about 1.0 in soft clay to below 0.4 in stiff clay as c_u/p_a rises. It represents the short-term, end-of-driving condition and its accuracy depends entirely on the quality of the undrained strength profile. The β method is an effective-stress method: f = β σ′_v with β = K tan δ′, and for a normally consolidated clay K ≈ K₀ = 1 − sin φ′_R. It represents the long-term drained condition after the excess pore pressures generated by installation have dissipated, and it is the physically correct model of what a shaft actually resists over the life of the structure. The λ method of Vijayvergiya and Focht is a hybrid: the average unit friction over the whole embedded length is f_av = λ(σ′_v(av) + 2c_u(av)), where λ is a single dimensionless coefficient that decreases with embedded length — 0.500 at the surface, 0.336 at 5 m, 0.245 at 10 m, and about 0.11 at 60 m.

Prefer the λ method when the pile is long and driven, and the profile is layered. The method was regressed from load tests on long, large-displacement steel pipe piles supporting offshore platforms in the Gulf of Mexico, and its natural domain is exactly that: driven displacement piles of substantial length in soft to medium normally consolidated or lightly overconsolidated marine clay. Its most useful property is that the length-dependence of λ captures, in one empirical coefficient, the well-documented observation that average unit friction does not grow in proportion to depth on a long pile — a shortening effect variously attributed to progressive failure along a compressible shaft, to stress relief as the pile advances, and to the reduction of lateral stress behind the advancing tip. Neither the α nor the β method contains any such length effect, so both tend to over-predict the capacity of a very long pile. In Canadian practice this makes the method attractive for long driven piles in the deep soft deposits of the St. Lawrence lowlands, the Fraser delta and the Beaufort shelf.

Prefer it also when the data are averages rather than profiles. Because λ is applied to depth-averaged values of σ′_v and c_u, the method needs only mean values over the embedment and does not require the layer-by-layer c_u profile that the α method consumes, nor the drained strength parameters that the β method needs and that a routine investigation of a soft clay often does not report. It is therefore a robust first estimate at preliminary design stage, and a good independent check on an α-method answer.

Prefer α when the clay is a single well-characterised layer and the pile is short. The α method uses the local strength at each depth, so it handles a strongly varying c_u profile, a crust over soft clay, or a short pile of ten metres or less far better than a single averaged coefficient can. It is also the method for which the largest body of load-test verification exists for driven piles in onshore clays, and the appropriate method for the end-of-driving condition that governs driving resistance and set-up predictions.

Prefer β whenever effective stress is the physical driver. That covers heavily overconsolidated or fissured clays, where a total-stress adhesion factor is unreliable; bored and augered piles, where installation does not generate the pore pressures the λ data set embodies; any long-term or drained assessment; and — decisively — every problem in which the vertical effective stress will change during the design life, such as negative skin friction from a new fill, a permanent change in the groundwater table, or a pile in a layered clay–sand profile where the sand must be treated in effective stress anyway. The β method also extends naturally into the sand layers of a mixed profile, which the other two do not.

In summary, use λ for long driven displacement piles in relatively uniform soft to medium normally consolidated clay, particularly offshore or in deep marine deposits, and as a fast independent check; use α for short or moderate-length driven piles in a well-defined c_u profile and for end-of-driving capacity; use β for bored piles, stiff or overconsolidated clays, drained and long-term conditions, and any profile in which effective stresses will change. Best practice, and the practice this paper's Question 6 asks for, is to compute the capacity by at least two of them and to reconcile the difference before adopting a design value.