A Consolidation Prediction That Overshot Field Settlement by a Factor of Five
A 6 m thick soft to medium clay is to be surcharged by 4 m of granular fill. An oedometer test on a 75 mm tube sample from 8 m depth gave e0 = 1.15, Cc = 0.38, and a Casagrande preconsolidation pressure of 130 kPa. The in-situ vertical effective stress is 95 kPa and the fill adds 60 kPa. The original estimate, made by applying Cc across the whole stress range, was 220 mm. Fourteen months after fill placement, the settlement plates read 45 mm and the rate has fallen to nearly zero. The fill is scheduled to be removed and the slab cast.
1. Recompute the consolidation settlement and explain why the original estimate was so high.
Read the worked answer
The stress increment straddles the preconsolidation pressure, so two slopes apply.
From 95 to 130 kPa the clay recompresses on Cr; from 130 to 155 kPa it follows the virgin line on Cc.
s = H/(1+e0) x [ Cr log(130/95) + Cc log(155/130) ]
- H = 6 m, 1+e0 = 2.15, so H/(1+e0) = 2.79 m
- Cr term: 0.06 x log(1.368) = 0.06 x 0.136 = 0.0082
- Cc term: 0.38 x log(1.192) = 0.38 x 0.076 = 0.0290
- Sum = 0.0372
- s = 2.79 x 0.0372 = 0.104 m, about 105 mm
The original 220 mm came from applying Cc from 95 kPa to 155 kPa without recognizing that most of the first increment is recompression. Using Cr for that portion removes roughly half the settlement.
2. The field reads 45 mm, not 105 mm. Is your recomputation confirmed? What would you check before approving the slab?
Read the worked answer
Not confirmed - the recomputation is closer but still 2x the measurement.
Possible explanations to check, in order of likelihood:
- Sample disturbance. A 75 mm tube sample in soft clay is disturbed enough that the Casagrande construction under-predicts the true preconsolidation pressure and Cc is biased high. A disturbed curve is flatter and the break is smeared. Re-examine the curve: if the compression index in the recompression range is nearly as steep as the virgin range, the sample is not usable for Cr.
- The real OCR is higher than 1.37. Check the geological history (erosion, desiccation, aging) and consider the strain-energy or void-ratio-index method to pick the preconsolidation pressure.
- Pore pressure. Has dissipation actually finished? "Rate has fallen to nearly zero" is not the same as u = 0. Read the piezometers; if excess pore pressure remains, settlement is incomplete and you must not remove the fill or cast the slab yet.
- Where the plates are. A plate over a thinner clay section, or partly on a stiff crust, will under-read the design section. Confirm the plate locations against the borings.
- Fill thickness and unit weight as actually placed may differ from the design 4 m.
3. What would you change about the sampling and testing for the remaining design, and how does that affect the numbers you would use?
Read the worked answer
Sampling and testing changes
- Replace tube sampling with 75 mm piston samples (or block samples) in the soft clay, and take them at the exact design depths, not at convenient intervals.
- Run a constant-rate-of-strain (CRS) oedometer test in addition to the incremental test; CRS gives a continuous curve and makes the preconsolidation break easier to identify.
- Test at least two samples per stratum and compare curves; a single oedometer curve is not a design basis.
- Use Bjerrum's correction for sample disturbance on the field curve, or backfigure Cc and Cr from the settlement plates plus measured pore pressures once dissipation is complete.
What to do with the numbers
- Use Cr for any stress path that stays below the preconsolidation pressure and Cc only above it. If the corrected preconsolidation pressure lands near 180 kPa, the entire fill load is in recompression and settlements drop to a few tens of millimeters.
- Do not use the field curve as a substitute for design at another location unless the soil profile at the plates matches the design section; use it to validate, not to replace, the lab data.
- Report the settlement estimate as a range with the assumptions stated, and keep the slab tolerant of the remaining movement (void former, structural slab, or delayed casting with a monitoring trigger).