Reading a geotechnical report without over-trusting the SPT correlations

Kavita YilmazRemote Sensing Engineering · 21 yrs

Good summary and the distinction between what was measured and what was derived is the one thing most design engineers do not carry across from the report into their model. One addition to the list of what to check first: how the SPT was actually performed, specifically the hammer energy ratio. Almost every published correlation is calibrated against a corrected N value at 60 percent energy, and an uncorrected field N from an automatic trip hammer at 80 percent energy fed into those correlations will give you a systematically unconservative stiffness. Whether the correction was applied is often buried in the methodology section or absent entirely, and it is a first-order effect rather than a refinement.

The related question worth asking of any report is whether the summary table values are the measured values or the interpreted design values, because reports vary and the difference can be substantial. An interpreted characteristic value already has a degree of conservatism applied by the geotechnical author, and applying your own partial factors on top of that double-counts. Conversely, taking an interpreted mean as though it were characteristic goes the other way. The table rarely says which it is and it is always worth one email.

Zara FaroukConstruction Engineering · 21 yrs

Verified expertise in Construction Engineering

The advice to ask for direct measurement rather than adding another safety factor is right and it is worth explaining the mechanism to whoever has to approve the cost. Adding a factor to a correlated stiffness does not reduce the uncertainty, it just moves your answer in one direction by an arbitrary amount — and for a stiffness-governed design the direction that is conservative is not always obvious, because a stiffer soil is unconservative for some checks and conservative for others. Dynamic soil-structure interaction and any frequency-sensitive design are the clear cases where a blanket conservative factor can put you in a worse position than the best estimate.

On which test to ask for, match it to the strain level your design actually operates at. A seismic cone or cross-hole gives small-strain stiffness, which is what you want for dynamic and serviceability stiffness checks. A pressuremeter gives you stiffness at an intermediate strain closer to the working range of a loaded foundation. A plate test gives near-surface behaviour at working strain over a small influence zone. They are not interchangeable, and asking for the wrong one is an expensive way to end up back where you started.

Sofia AdeyemiMine Automation Engineering · 12 yrs

Your point about borehole depth relative to the influence zone is the one most frequently violated in my review experience, and the pattern is consistent: the SI was scoped before the foundation concept was settled, so the depths were chosen against a generic assumption. If the concept later moves from pads to a raft or to piles, the influence zone changes completely and the existing boreholes may not reach it. The report is then perfectly valid for a foundation nobody is building.

The habit that avoids it is to sketch the influence zone for the actual foundation concept on the borehole location plan before reading anything else. It takes ten minutes and it immediately shows whether the data supports the design, whether coverage is adequate laterally as well as vertically, and whether the variation between boreholes is being averaged across a distance that is too large to justify. If the zone extends below the deepest hole, that is a scope conversation to have before you start calculating, not after.