16-Civ-B3 Geotechnical Design · December 2016

Question 5 of 9: When to prefer the CPT over the SPT

Nivaar worked solution (AI-drafted; not reviewed by a licensed engineer)

Notes on this paper

Paper format. Professional Engineers Ontario / Engineers Canada National Examinations, December 2016 — 98-Civ-B3 Geotechnical Design. Three hours, OPEN BOOK, any non-communicating calculator. Section A carries five discussion questions of 7 marks each (answer any four); Section B carries four design questions of 24 marks each (answer any three); the examinable total is 4 × 7 + 3 × 24 = 100 marks. All nine questions are worked below, because the set is a study resource rather than a timed attempt.

Reference texts (98-Civ-B3 / 16-Civ-B3 Geotechnical Design).

B. M. Das, Principles of Foundation Engineering, 9th ed. — subsurface exploration (Ch. 2), bearing capacity of shallow foundations (Ch. 3), settlement of shallow foundations including Schmertmann's strain-influence method (Ch. 5), retaining walls (Ch. 8), pile foundations (Ch. 11), drilled shafts (Ch. 12).
B. M. Das, Principles of Geotechnical Engineering, 9th ed. — shear strength (Ch. 12), lateral earth pressure (Ch. 13), slope stability (Ch. 15).
Canadian Geotechnical Society, Canadian Foundation Engineering Manual (CFEM), 4th ed. — the governing Canadian practice document for site investigation, in-situ testing, bearing resistance, deep foundations and earth-retaining structures.
R. F. Craig, Craig's Soil Mechanics, 9th ed. — effective stress, undrained versus drained behaviour, earth pressure, and the total-stress circular-arc slope analysis.
M. J. Tomlinson & J. Woodward, Pile Design and Construction Practice, 6th ed. — shaft adhesion in clay and under-reamed (belled) bored piles.
P. K. Robertson & K. L. Cabal, Guide to Cone Penetration Testing, 6th ed. — CPT interpretation and CPT-versus-SPT selection.

Sources of design charts and assumed values (page-1 Note 6). Note 6 requires the candidate to identify the source of every design chart and of every value assumed where the paper supplies none. They are named at the point of use and collected here:

Adhesion factor α = 0.55 for a drilled shaft in clay (Q6) — O'Neill and Reese (1999), reproduced in Das, Principles of Foundation Engineering, 9th ed., Section 12.9; the exclusion of the top 1.5 m and of one shaft diameter above the bell comes from the same source.
Bearing-capacity factor N_c^* = 9 (Q6) — Skempton (1951), as tabulated in Das, Section 11.11.
Overburden correction C_N = √(p_a/σ'_o) (Q7) — Liao and Whitman (1986), Das Principles of Geotechnical Engineering, Section 17.6.
SPT-to-friction-angle correlation (Q7) — Peck, Hanson and Thornburn (1974) as fitted by Wolff (1989); cross-checked against Kulhawy and Mayne (1990). Both are tabulated in Das, Principles of Foundation Engineering, 9th ed., Section 2.9.
Bearing-capacity factors and shape/depth factors (Q7) — Vesic (1973) and De Beer (1970), Das Sections 3.6 and 3.7.
Strain-influence diagram and the C₁, C₂ factors (Q7) — Schmertmann, Hartman and Brown (1978), Das Section 5.6; the modulus correlation E_s = 500(N₆₀ + 15) kPa is Bowles (1996), reproduced in the same section.
Rankine active coefficient for an inclined backfill (Q9) — Das, Principles of Geotechnical Engineering, 9th ed., Eq. (13.35); the base friction and adhesion reductions k₁ = k₂ = 2/3 are Das Section 8.4.
Unit weight of the submerged backfill (Q9) — assumed equal to the printed moist unit weight, 18 kN/m³, in the absence of a saturated value; the consequence of that assumption is bounded in the Q9 callout.

Section A

Question 5: When to prefer the CPT over the SPT (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.

The cone penetration test is preferred whenever the design question turns on continuity, repeatability, or pore pressure, and the ground is soft enough to push a cone into. The standard penetration test survives in practice because it recovers a sample and because it works in ground the cone cannot enter. Stated as a preference rather than as a ranking, the cases divide as follows.

Prefer the CPT when the profile matters. The cone records tip resistance, sleeve friction and pore pressure every 10 to 20 mm, giving a continuous record; the SPT gives one number per 1.5 m of borehole and tells you nothing about the 1.05 m between blows. Interbedded and varved deposits, a firm crust over soft clay, a thin loose seam in an otherwise dense sand, or a thin drainage layer inside a clay are all routinely missed by the SPT and unmistakable in a CPT trace. Where the design is controlled by the weakest thin layer — slope stability, liquefaction, punching through a crust — that is decisive.

Prefer the CPT in soft to firm clays and silts and in loose to medium sands. In these materials SPT N values are small integers, so the resolution is poor and the scatter is large; cone resistance is a continuous variable with good precision. Undrained shear strength follows from c_u = (q_t − σ_vo) / N_kt, and the piezocone's u₂ reading additionally gives the in-situ pore pressure directly, identifies drainage boundaries, and — through a dissipation test — yields the horizontal coefficient of consolidation, a parameter the SPT cannot supply at all.

Prefer the CPT when repeatability and operator independence matter. The SPT is notoriously sensitive to hammer type and energy ratio, rod length, borehole diameter, sampler liner, and the care with which the hole is cleaned; the raw N must be corrected to N₆₀ and then for overburden before it means anything. The cone is a calibrated electrical instrument pushed at a standard 20 mm/s, and its output is comparable between rigs, operators and countries. For a project where two contractors' data must be combined, or where the same site will be re-investigated years later, that reproducibility is worth a great deal.

Prefer the CPT for soil behaviour classification, liquefaction assessment, and direct design methods. Robertson's normalised soil-behaviour-type charts classify the ground from q_t, F_r and B_q without a sample. Canadian and international liquefaction practice is now anchored on q_c1N-based cyclic resistance ratios, with the SPT-based curves retained mainly for legacy comparison. Pile design by direct CPT methods (LCPC, Eslami-Fellenius, and the CFEM's own procedures) uses cone resistance as the input and is generally more reliable than N-based pile formulae.

Prefer the CPT on grounds of speed, economy and disturbance. A cone rig commonly achieves 100 to 200 m of sounding per day against perhaps 30 m of drilled and sampled borehole; no spoil is generated, which matters on contaminated sites; and in sensitive clays the absence of a drilled hole avoids disturbing the very material being tested.

Retain the SPT where the CPT cannot go or cannot answer. Gravels, cobbles, boulders, glacial till with clasts, fill containing rubble, cemented soils, and weathered rock will refuse a cone or damage it, whereas the SPT sampler can be driven and, if necessary, the hole advanced by coring. The SPT also recovers a disturbed sample, which is often the only direct evidence of grain size, colour, odour and fabric — and identification of the soil is not a small matter. Finally, a very large body of empirical correlation is expressed in N₆₀, so on small conventional projects where an N-based allowable pressure is entirely adequate, the SPT remains the economical choice. Best practice on a significant project is not to choose but to combine: CPTu soundings for profiling and parameters, with a smaller number of sampled boreholes to identify the materials and to calibrate the cone correlations against laboratory tests.

Topic. Selection between the cone penetration test and the standard penetration test in routine practice.

Key relations. CPT: $c_u = \dfrac{q_t - \sigma_{vo}}{N_{kt}}$ with $N_{kt}$ typically 12 to 18; $q_t = q_c + u_2(1-a)$; normalised tip resistance $Q_{tn} = \left(\dfrac{q_t-\sigma_{vo}}{p_a}\right)\left(\dfrac{p_a}{\sigma'_{vo}}\right)^{n}$. SPT: $N_{60} = \dfrac{N \cdot E_r \cdot C_B \cdot C_S \cdot C_R}{60}$ and $(N_1)_{60} = C_N N_{60}$.

Why this works. The two tests measure different things. The cone measures a continuous resistance under a controlled, quasi-static, known boundary condition, so its output is a soil property modified by a well-understood geometry. The SPT measures the dynamic resistance of a thick-walled sampler driven by a hammer whose delivered energy varies from rig to rig, so its output is a soil property convolved with the equipment. Wherever the design needs a parameter rather than an index, the cone wins; wherever it needs a sample or must penetrate coarse ground, the SPT wins.

Common pitfall. Treating the choice as absolute. Recommending a CPT-only programme on a till site guarantees early refusal and a wasted mobilisation; recommending an SPT-only programme in a soft varved clay guarantees that the layer which controls the design is never seen. A second pitfall is comparing raw N with q_c without correcting N to N₆₀ first.

Source. Robertson and Cabal, Guide to Cone Penetration Testing, 6th ed., Sections 2 and 4; Das, Principles of Foundation Engineering, 9th ed., Sections 2.9 (SPT and its corrections) and 2.11 (CPT); CFEM, 4th ed., Ch. 5.

← Question 4 Next: Question 6 →