Low Cylinder Breaks and a Request to Edit the Test Report
A 120 cubic metre pour of a 28 MPa specified compressive strength reinforced concrete mat was sampled and tested. The 28-day results, where each strength test is the average of two 150 x 300 mm cylinders tested to ASTM C39, are: set 1, 26.4 and 25.8 MPa for a test result of 26.1; set 2, 24.2 and 25.0 MPa for 24.6; set 3, 25.6 and 26.0 MPa for 25.8.
The drawings specify 28 MPa at 28 days and the mix design had a target mean strength of 34 MPa. The laboratory's own coefficient of variation on this mix is 3.5 percent. The project manager wants the report issued tomorrow marked pass, arguing the results are within 10 percent and that core testing would delay structural steel erection. The PM has asked the technician to reissue the report with the two lowest cylinders excluded because the ends were 'probably not ground properly.' There is no record of specimen damage or end-preparation problems, and the cylinders were logged in and broken by the laboratory in the normal way.
1. Apply the ACI 318 acceptance criteria to these results and state whether the concrete is acceptable.
Read the worked answer
Two criteria apply to a strength test result, which is the average of two cylinders. First, the average of any three consecutive strength tests must equal or exceed the specified strength. Second, no single test may fall below the specified strength minus 3.5 MPa.
Average of the three tests: (26.1 + 24.6 + 25.8) / 3 = 25.5 MPa against 28 MPa required, so the first criterion fails by 2.5 MPa, roughly 9 percent. The lowest single test, 24.6 MPa, sits just above the 24.5 MPa floor, so the second criterion passes narrowly. The lot does not meet the acceptance criteria.
The low coefficient of variation matters to the interpretation. With a laboratory COV of 3.5 percent and a target mean of 34 MPa, a 24.6 MPa result is about eight standard deviations below the mean. That is not a plausible cylinder-preparation artefact; it points to a genuine shortfall in the delivered concrete or to a sampling problem. Attributing it to grinding is not supported by the data.
The correct next step is core testing as provided by the code: at least three cores per lot taken per ASTM C42, tested wet, where acceptance requires the core average to be at least 0.85 times the specified strength (23.8 MPa) and no single core below 0.75 times (21.0 MPa). If the cores comply, the lot is acceptable without further evaluation.
2. How do you respond to the PM's instruction to exclude the low cylinders and mark the report pass?
Read the worked answer
Refuse, in writing and immediately. The test results are the record of what happened; excluding valid data after seeing the outcome and reissuing the report as a pass is falsification, regardless of the intent behind it. It would also breach the laboratory's accreditation requirements, which require records to be complete and unaltered, and the professional obligation to report findings accurately.
The narrow exception is a specimen shown to be defective on documented grounds, for example damaged in transit, improperly end-ground with the defect recorded at the time, or containing a void visible at fracture. That record must exist independently of the result. Here there is none, and the COV argument shows the low breaks are consistent with real concrete, not with bad specimens.
What I would do instead: issue the report with all results, the cylinder identification, sampling locations, break values, the criteria applied and the conclusion that the lot fails. Then advise the PM of the code-compliant route, which is core testing, and give the realistic duration for coring, testing and reporting so the steel erection decision can be made on facts. Offer to attend the coring to witness locations and repairs.
If the PM insists on altering the report, escalate to the laboratory's quality manager and the engineer of record, and record the instruction and my refusal in the project file.
3. Assume the cores return 25.4, 24.9 and 26.3 MPa. What do you conclude, and what else must be checked before the mat is signed off?
Read the worked answer
The core average is (25.4 + 24.9 + 26.3) / 3 = 25.5 MPa, above the 23.8 MPa threshold, and the lowest core at 24.9 MPa is well above the 21.0 MPa floor. On the code criteria, the lot is acceptable and no further evaluation of the concrete as a material is required.
Acceptance is not the end of the engineering. The measured strength is about 9 percent below the specified value, and the structural design was based on the specified value. The engineer of record must be informed and must confirm that the capacity checks still hold with the lower in-place strength: flexure is normally steel-governed and may be unaffected, but shear, punching, development and anchorage, and serviceability should be re-checked. If the shortfall had been larger, or the cores had failed the criteria, the route would be a structural evaluation, and possibly a load test under the code's existing-structure provisions.
Administratively: record the core locations, the take and test dates, the repair of the core holes, the engineer of record's written response, and the original failing cylinder results. The failed data stays in the record; it is part of the concrete's history and is relevant to any future investigation or durability assessment.