16-Civ-B3 Geotechnical Design · Undated paper
Nivaar worked solution (AI-drafted; not reviewed by a licensed engineer)
Paper format. National Examinations (Engineers Canada / EGBC), 16-Civ-B3 Geotechnical Design — 3 hours, open book, any non-communicating calculator permitted (the candidate must write its make and model on the left-hand sheet). The paper prints nine questions in two sections: Section A holds five short questions of 7 marks and asks for any four; Section B holds four design questions of 24 marks and asks for any three. Only the first four of Section A and the first three of Section B are marked, so a complete paper is 4 × 7 + 3 × 24 = 100 marks. Note 1 urges the candidate to state any assumptions made, Note 6 requires the source of every design chart to be identified, and Note 7 permits assumed values provided the source is stated. All nine questions are solved below, because the set is a study resource rather than a sitting.
Reference texts. B. M. Das, Principles of Foundation Engineering, 8th ed. (bearing capacity ch. 3, settlement of shallow foundations ch. 5, drilled shafts ch. 12, retaining walls ch. 8, sheet pile walls ch. 9); B. M. Das, Principles of Geotechnical Engineering, 9th ed. (shear strength, lateral earth pressure, slope stability); R. F. Craig and J. A. Knappett, Craig’s Soil Mechanics, 8th ed. (effective stress, undrained strength, anchored walls); D. P. Coduto, Foundation Design: Principles and Practices, 2nd ed.; and in the Canadian frame the Canadian Foundation Engineering Manual (CFEM), 4th ed., Canadian Geotechnical Society — ch. 4 for site investigation and in-situ testing, ch. 10 for shallow foundations, ch. 18 for deep foundations and ch. 25 for earth retaining structures. Test standards are quoted as ASTM/CSA where the CFEM adopts them (SPT: ASTM D1586; CPT: ASTM D5778; field vane: ASTM D2573).
Source-quality note.
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.
[Figure not reproduced: The equipment offered in Figure 1 of the paper, redrawn, with the design quantity each one actually delivers. Only the plate load test loads the soil the way a footing does; only the CPT profiles it continuously. See the official exam paper.]
Recommendation. The electric cone penetration test with pore-pressure measurement (piezocone, CPTu), run to at least twice the width of the largest proposed footing below founding level, with one companion sampled borehole for classification. In the words of the question, it is the single piece of equipment that supplies the parameters a rational design of a shallow foundation on sand actually needs.
Why the CPT, and not one of the others. The controlling fact about sand is that it cannot be sampled undisturbed at any sensible cost: a tube sample of clean sand loses its fabric, its in-situ density and its stress history the moment it is recovered, so laboratory strength and stiffness testing on such a sample is not design evidence. Every parameter must therefore be measured in place. The cone measures tip resistance $q_c$, sleeve friction $f_s$ and pore pressure $u_2$ continuously at 10 to 20 mm intervals, so a loose seam 300 mm thick is seen; the standard penetration test, sampling at 1.5 m centres, can step straight over it. Because the cone is pushed hydraulically at a controlled 20 mm/s and read electronically, the result does not depend on the driller: SPT blow counts vary with hammer energy, rod length, borehole diameter and whether the hole was kept full of mud, which is why they must be corrected to $N_{60}$ before they mean anything. For design the cone feeds the whole chain directly — relative density and friction angle from $q_{c1}$, constrained modulus and the Schmertmann strain-influence settlement method from $q_c$, soil behaviour type from the $q_t$–$F_r$ chart, the position of the water table and any silt or clay interbeds from $u_2$ and its dissipation, and, in seismic British Columbia, the cyclic resistance ratio for liquefaction screening from $q_{c1N,cs}$. It is also fast: 60 m of continuous profile in a working day, at a fraction of the cost of the equivalent length of sampled borehole.
The alternatives in Figure 1 each answer a narrower question. The SPT is cheap and universal and does return a sample, but it is discontinuous and operator-sensitive. The field vane measures undrained strength and is meaningless in a free-draining sand. The pressuremeter gives a genuine in-situ stress–strain curve and therefore the best modulus of any of them, but it needs a pocket to be prepared, is slow, and gives one test per metre at best. The plate load test loads the ground exactly as a footing does, which is its appeal, but a 300 mm plate only stresses about 600 mm of depth, so it says nothing about a compressible layer beneath a 3 m pad — the classic reason plate load tests mislead on sand. Geophysics gives stratigraphy and small-strain stiffness over large areas, but not the design bearing parameters.
Limitations, which must be designed around. First, the cone recovers no sample: there is no grading curve, no fines content, no plasticity, no chemical or contamination screening, and soil type is inferred from a behaviour-type chart rather than seen. The investigation must therefore pair the cone soundings with at least one sampled borehole, and the CFEM expects exactly this pairing. Second, the cone refuses in gravel, cobbles, boulders, cemented horizons and very dense sand: on a 20 t rig, refusal in dense glacial material at 10 to 15 m is routine, and on a site with a coarse fill or a till cap the cone may not reach founding level at all. Third, it needs a heavy reaction mass and firm level access, which rules out steep, soft or confined sites. Fourth, every design parameter comes from an empirical correlation calibrated largely on clean, uncemented, unaged quartz sands; aged, cemented, calcareous or micaceous sands, and sands with appreciable fines, will be misread unless local calibration exists. Fifth, at depth the rods can deviate from vertical, and in partially drained silty sands the measured $q_c$ is neither drained nor undrained, so the correlations sit in a grey zone.
