Five Low Strength Tests on 35 MPa Columns: Validity and Core Route
The columns of a 12-storey frame are specified as 35 MPa concrete. Five strength tests, each the average of two 150 × 300 mm cylinders, returned 37.1, 34.1, 32.3, 35.3 and 30.6 MPa at 28 days. The cylinders were cast by the supplier's technician and left in a site container for three days, where afternoon temperatures reached 34 °C, before being transferred to a curing tank; they were capped with sulfur mortar about four hours before testing. One cylinder in the fifth test fractured on a diagonal plane running from the cap instead of forming the usual cones. The contractor argues that all results are within 10% of the specified strength, that the 30.6 MPa pair is an outlier, and that the columns should be accepted without further investigation. The building is in a non-aggressive environment and the specified maximum w/c is 0.45.
1. Apply the acceptance criteria to these five tests and state which criteria fail.
Read the worked answer
Criteria (f'c = 35 MPa). Every average of three consecutive strength tests must be at least f'c; no individual strength test may fall below f'c − 3.5 = 31.5 MPa.
Three-test averages.
- Tests 1–3: (37.1 + 34.1 + 32.3)/3 = 34.5 MPa - below 35. Fails.
- Tests 2–4: (34.1 + 32.3 + 35.3)/3 = 33.9 MPa - fails.
- Tests 3–5: (32.3 + 35.3 + 30.6)/3 = 32.7 MPa - fails.
Individual tests. Test 5 at 30.6 MPa is below 31.5 MPa - fails.
Result. Two criteria are violated, not one. Note that the first three-test average failing by only 0.5 MPa is a real failure, not a rounding issue: the criteria are absolute thresholds, and the trend is downward across the pour rather than random scatter, which is what usually signals a genuine problem (cement change, water, or curing) rather than testing noise.
2. Evaluate the contractor's 'within 10%' and 'outlier' arguments, and assess how the cylinder handling affects the numbers.
Read the worked answer
The 10% argument. There is no 10% tolerance in the acceptance criteria; conformity is judged against f'c and f'c − 3.5 MPa. A result band around the target is not a specification.
The outlier argument. 30.6 MPa is already the average of a pair, so the criterion is applied to it directly. Discarding it because it is inconvenient would also leave only four tests.
The curing defect is real, and it cuts the other way. For acceptance, cylinders must be standard-cured at 20 ± 2 °C from casting to test. Three days at up to 34 °C accelerates early hydration but typically depresses the 28-day strength, so the reported values may understate the concrete in the columns. That is a legitimate reason to doubt the numbers - but it is a reason to test the structure, not to declare it compliant.
The diagonal fracture. A shear-plane failure through the cap usually indicates a capping or platen problem; record it and check whether the companion cylinder agrees. Capping four hours before testing is acceptable only if the sulfur has fully hardened and the caps are parallel.
3. Design the follow-up investigation and state the acceptance criteria and the consequence of failing them.
Read the worked answer
Programme. Take at least three cores, 70–100 mm diameter, from locations that represent the concrete in question - distributed across the affected columns and floors, avoiding reinforcement and embedded items, and not taken from cosmetic patches. Moisture-condition them and test in a moist condition, and report density and any visible defects (voids, laitance) alongside strength.
Criteria. The average of three cores must be at least 0.85 f'c = 29.75 MPa, and no single core may be less than 0.75 f'c = 26.25 MPa.
Interpretation. Cores are cast and cured under site conditions, so they measure the concrete as it actually is; a core set meeting 0.85 f'c supports acceptance of the member, subject to the engineer confirming that the in-place strength is adequate for the loads and that durability is unaffected. If the cores fail, the structure is evaluated by analysis, and options include strengthening, load restriction, or removal of the affected elements.
Do not skip the durability question. Even if cores pass, confirm that the w/c actually delivered did not exceed the 0.45 limit; in a non-aggressive environment this is less critical, but the record must show it was checked.