Concrete Conformity Failure on a Three-Storey Frame

The specification is for a characteristic strength, not an average. C30/37 means a characteristic cube strength of 30 MPa with a 5 percent failure probability. Conformity is judged on production, not on a single set.

Take the identity criteria applied in common practice for a production period (for example EN 206, group 1):

• Mean of the results: fcm must be at least fck + 4 = 34 MPa.
• Any individual result: at least fck − 4 = 26 MPa.

Apply to the data.

• Mean 32.5 MPa is below 34 MPa. The contractor says 32.5 is above 30; it is not above the 34 required for the mean, so the argument fails on its own terms.
• Lowest result 24.0 MPa is below 26 MPa, so the individual criterion also fails.
• Three results below 27 MPa, in a set of fifteen, indicate a skewed distribution, not random scatter.

So the concrete fails conformity on both criteria. The average is irrelevant to the question of whether the concrete as supplied belongs to the specified grade.

Other factors that must not be brushed aside

• Re-tempering 4 m3 with water at 3.5 hours raises the water/cement ratio and is not permitted without the engineer's written approval; it is the single most plausible explanation of the low results.
• Three days of curing is short for a hot climate, although it does not affect the 28-day standard cubes if they were cured correctly. It does affect the concrete in the structure.

The non-conformance report is correct. The contractor's proposal to proceed is premature: the upper floor adds load to the columns whose concrete is in question.

The objective is not to repeat the cube test in-situ, but to decide whether the structure can carry its loads. Standard cubes test the concrete delivered; cores test the structure.

Sampling

• Take at least three cores from the elements in question, preferably more, distributed across the pour and across columns and beams rather than clustered in one member. If the laboratory cannot give a confident answer from three, take six.
• Avoid beam-column joints, the anchorage zones, and any position where cutting a core would remove cover from the main reinforcement or cut the reinforcement itself. Use a covermeter to locate reinforcement before setting the rig.
• Nominal diameter 100 mm where the element permits, at least 75 mm otherwise; the length should be at least one diameter after the ends are cut back, with the axis perpendicular to the concrete surface.
• Test in compression with the core ends prepared flat, and if the core contains reinforcement, either exclude it or account for it in the interpretation.

Complement with non-destructive testing

• Ultrasonic pulse velocity and rebound hammer on the suspect columns, calibrated against the cores. Their value is in mapping uniformity across the pour, not in giving an absolute strength; several calibrating cores are required before the readings are used for a decision.

Interpretation

• Convert the core results to an in-situ cube strength using the published core-to-cube relationship in the relevant standard, including the length/diameter factor and any correction for the presence of reinforcement.
• Compare against the strength the structure actually needs at the design stage of construction. A structure that needs only, say, 70 percent of its 28-day strength at the time of loading may well pass; that is a structural check, not a conformity check.
• If the cores confirm at least the design-required strength, the element can be accepted on the basis of a structural check, with the non-conformance recorded and the durability implications (cover, curing, permeability) considered separately.
• If not, the options are local strengthening, section enlargement or replacement of the affected columns, and a revised construction sequence.

Timing: cores can be tested within about a week of extraction. There is no engineering reason to let the upper floor proceed before the results are in.

Root cause analysis

The pattern is consistent with a combination of:

1. Over-retardation. A retarding admixture in 34 degrees Celsius heat is applied to slow setting, but the last truck was 3.5 hours old at discharge. Retarder extends setting time from the moment of batching, not from arrival. A retarded mix placed at 3.5 hours may still be workable but may already be losing early strength development, and it is the late-delivered concrete that usually shows the low results.
2. Re-tempering with water. Adding water to restore slump from 60 to 110 mm raises the water/cement ratio directly. A slump increase of that magnitude is not achieved by a few litres; it is a real strength reduction, and it is the single largest candidate cause of the 24 MPa result.
3. Delivery timing and site organisation. The pour finishing at 22:30 with a hot mix and slow discharge points to insufficient truck spacing and poor planning, not to a material defect.
4. Curing. Three days of wet hessian is inadequate for a hot climate; typical good practice is seven days. It will not change the standard cube results but will affect the surface zone, cover quality and long-term durability.

Decision on the upper floor

Prohibit construction of the next floor on the affected columns until the core results are available. The reason is not procedural caution: the upper floor adds axial load and moment to exactly the elements whose strength is unverified, and if the concrete fails you will be testing a structure that is already carrying two floors of unverified load.

Practical programme

1. Cores this week; results within ten days.
2. In parallel, obtain the batching plant records for every truck (time, slump, temperature, admixture dosage) so the cause is documented, not guessed.
3. Keep the non-conformance report open.
4. If the cores pass, permit the next floor with the engineer's written acceptance, and tighten the concrete specification: no site re-tempering without written approval, a maximum delivery time related to the ambient temperature, mandatory slump and temperature records at the point of discharge, and a curing requirement of seven days.
5. Back-charge the investigation to the contractor if the cause proves to be site practice, as the records are likely to show.