A two-storey commercial building goes up on a low-lying site along the Bulacan or Pampanga coast, the Cavite shoreline, or on reclaimed ground in Cebu or Iloilo. The soil report gives an allowable bearing capacity of 150 kPa at 1.5 m depth. The footings are sized to that number, they check out, and the building is turned over clean.
About three and a half years later the owner calls. Diagonal cracks have opened at partition corners next to the heavier interior columns, doors near the middle bay are binding, and a level run shows the ground-floor slab sloping toward the centre of the building. By then the clay is essentially at the end of primary consolidation, but the cracking started long before the call: roughly 78 mm was already in the ground at the twenty-month mark. The footings were never overstressed. The soil under them simply took years to squeeze out its water.
The check that would have caught it is a one-dimensional consolidation settlement estimate, done by hand from the soil report in about fifteen minutes.
The most misread number in a Philippine soil report is the allowable bearing capacity. In most reports it is an ultimate bearing capacity from a shear-strength calculation divided by a factor of safety that runs about 2.5 to 3.0 in practice, varying with the quality of the data behind it. It says the footing will not punch through the soil. Often it says nothing about how far the footing will move, but not always: competent reports frequently quote the allowable as the lesser of a shear-based value and a settlement-limited value, commonly on a 25 mm criterion, and then the number already carries a settlement check. Find out which criterion governed, and for what footing width and founding depth, before assuming it is shear only.
NSCP 2015 leans on investigated soil data wherever a foundation investigation is required. It also permits presumptive allowable foundation pressures for ordinary structures, and that is the legitimate route by which a settlement calculation gets skipped altogether. The point worth holding on to is that a presumptive bearing value is explicitly not a settlement statement, and the code does not do the calculation for you.
All inputs are illustrative and typical of a low-lying Luzon site.
Given. Interior column, service load P = 600 kN unfactored, split 70 percent dead and 30 percent live, so with half the live load treated as permanent the sustained load is 420 + 90 = 510 kN. Footing 2.0 m × 2.0 m, base at 1.5 m. Stratigraphy: 0 to 3.0 m silty sand fill, γ = 18 kN/m³ moist and 20 kN/m³ saturated; groundwater at 1.5 m; 3.0 to 7.0 m soft gray clay, γsat = 17 kN/m³, e0 = 1.10, Cc = 0.35, LL = 50, cv = 2.0 m²/year; dense sand below 7.0 m, so the clay drains top and bottom. The clay is normally consolidated per the consolidation test, so σ'p equals σ'0 = 56.67 kPa. No site-raising fill is included. Immediate undrained settlement is excluded. cv is treated as constant.
The bearing check and what it rests on. Net pressure P/A = 600 ÷ 4.0 = 150 kPa, equal to the reported qa of 150 kPa, so bearing passes. That is like for like only because the report states its 150 kPa as a net allowable at founding level. A soil report's allowable at a stated founding depth is conventionally a gross value, and read that way the applied pressure is 150 + 18 × 1.5 = 177 kPa and this footing does not pass at all. If the report does not say which it is, ask, because the two readings land on opposite sides of the same check.
What those assumptions cost if they move. A desiccated crust or old fill commonly leaves σ'p 15 to 30 kPa above σ'0, putting part of the load on the recompression branch and roughly halving the settlement. One metre of site fill would add about 18 kPa over the whole footprint and take the single-slice settlement to about 148 mm. Immediate undrained settlement commonly runs 10 to 20 percent of the consolidation settlement for soft normally consolidated clay. And oedometer cv commonly varies by a factor of 2 to 3 with stress level, while field values are often several times higher where sand lenses drain horizontally, so the time figures below are the softest numbers here.
Skempton cross-check: Cc = 0.009 × (50 − 10) = 0.009 × 40 = 0.36, against the tested 0.35. That is not corroboration. The correlation is stated for normally consolidated clays of low to moderate sensitivity and carries roughly ±30 percent scatter, so at a liquid limit of 50 it only brackets roughly 0.25 to 0.47. Agreement to 3 percent is coincidence; the correlation confirms only that the tested value sits in a plausible band. Use the tested value.
Layer 3.0 to 7.0 m, H = 4.0 m, mid-depth 5.0 m, double drainage.
z = 5.0 − 1.5 = 3.5 m below the base. Δσ = 600 ÷ ((2.0 + 3.5) × (2.0 + 3.5)) = 600 ÷ 30.25 = 19.83 kPa.
Note which way that error runs. At depth-to-width ratios of roughly 1 to 2 the 2:1 method under-predicts the centre-line stress. For the upper sublayer used below, B = 2.0 m and z = 2.5 m, Boussinesq gives about 36 kPa against the 2:1 method's 29.6 kPa, roughly 20 percent low, and the two converge only for z greater than about 2B. The 2:1 answer is therefore unconservative for a serviceability check at these depths.
Compare with the 25 mm many designers use as a default screening limit for isolated footings, a rule of thumb traced to Terzaghi and Peck and originally meant for sand. For footings on clay the commonly cited figure is roughly 65 mm, the maximum total settlement for isolated foundations on clay from Skempton and MacDonald (1956). Neither is a code value, and 87 mm is past both.
Split the clay into two 2.0 m sublayers, each with Cc H ÷ (1 + e0) = 0.35 × 2.0 ÷ 2.10 = 0.3333 m.
The single slice was about 10 mm low. Keep 87 mm as the screening figure, since it is also what this building did, but 97 mm is the value to design against, and every decision below is rebuilt on it. A screening number that happens to match the observed damage is not a licence to keep using it once you know it under-predicts.
The single slice lands at 64.75 mm, a hair under the 65 mm figure, and that is the trap. On the refined basis it is 71 mm and still does not clear the limit: enlarging the footing alone did not solve the problem. Area up 125 percent (4.0 to 9.0 m²); settlement down about 27 percent (97 to 71 mm). Extra concrete at 0.50 m thick: (9.0 − 4.0) × 0.50 = 2.5 m³ per footing. At an illustrative in-place rate of ₱7,000/m³, that is 2.5 × 7,000 = ₱17,500 per footing (illustrative), before extra rebar and excavation, to buy 26 mm and still miss. The same 600 kN still reaches the clay; the wider base only spreads it a little.
The adjacent exterior footing carries P = 250 kN on 1.5 m × 1.5 m at the same depth, 6.0 m away.
1/142 is worse than the 1/150 the single-slice pair would have handed you, which is the whole argument for rebuilding the decision numbers on the refined settlements. The limits come from two different studies and deserve separate credit. Skempton and MacDonald (1956) put cracking of walls and partitions at about 1/300 and structural damage becoming likely at about 1/150; Bjerrum (1963) put the safe limit, where cracking is not permissible, at about 1/500. These are empirical observations, not code limits, and they were derived for framed buildings with panel infill. Load-bearing masonry and long walls are governed by deflection ratio rather than angular distortion and are more sensitive. At 1/142 the calculation has gone past the structural-damage threshold, which is exactly the cracked-partition, sloping-slab picture from the opening.
The clock starts when the load goes on. Part of the settlement happened before there were finishes to show it; most of the rest arrived after the partitions, tiles and door frames did, and the damage was visible for a long stretch before anyone counted it as a defect.
Immediate (undrained) settlement and secondary compression are excluded, and both add to the totals. The 2:1 spread and the net-pressure simplification are approximations, and as shown above the 2:1 error runs in the unconservative direction at these depths. Site fill and the overlapping stress from other footings are not included and can dominate on a raised site. All of these push the real number up.
Several things push it down, and overconsolidation is only one of them. What the oedometer gives you is an uncorrected one-dimensional settlement. A Skempton-Bjerrum correction for the three-dimensional stress path under a real footing typically reduces it by 0 to 30 percent for normally consolidated clay, with μ roughly 0.7 to 1.0. Sample disturbance that flattens the compression curve also reduces the computed value, and overconsolidation reduces it further, which is why σ'p still matters more than almost any other line in the report. The size of the effect is the point: applying μ = 0.8 takes the refined 97 mm to about 78 mm, and the enlarged 3.0 m footing from 71 mm to about 57 mm, the other side of the 65 mm figure. A correction that can move the answer across a decision boundary has to be a stated decision, not a silent omission.
The load basis works the same way. Using the sustained 510 kN instead of the full 600 kN takes the single-slice settlement to about 75 mm. That is another line the calculation has to declare rather than assume.
Enlarging the footing does not remove the load from the clay; it spreads the stress a little wider and deeper, and on the refined numbers it did not get this building under the limit. Once the check shows settlement of this order on normally consolidated clay, the realistic options are:
Which option wins is a cost-and-programme decision. The RHCES Estimator can price the extra concrete for enlarged footings, or the difference between footings and a raft, in a minute. It cannot tell you whether the bigger footing works. This check does, and the owner is paying for both.
Screen with what you have: unit weight and moisture content from the log, Cc from Skempton's correlation on the liquid limit, and e0 = w × Gs if the clay is saturated. Label everything an estimate, and remember the correlation only brackets a band. Ask whether the quoted qa is gross or net, and whether shear or settlement governed it. If the screening number lands anywhere near the tolerances, ask for undisturbed samples and a consolidation test before finalising the scheme. The test costs a small fraction of one enlarged footing.
A little, and less than intuition says. The example more than doubled the area and trimmed the refined 97 mm to 71 mm, which is still past the 65 mm figure. The stress reaching the clay depends on total load and depth, not only on contact pressure. Bigger footings are a bearing-capacity fix; on deep soft clay they are a poor settlement fix.
Often yes. A deeper layer sees a smaller increase from each footing, but stress from raised site fill does not reduce with depth, and adjacent footings' bulbs merge into something close to a uniform load down there. Run the same check with the deeper mid-depth and include the fill; a thick soft layer at 6 to 12 m under a fully loaded site can still settle more than the finishes can take.