Sand Moisture Probe Under-Reads: Tracing a 28-Day Strength Failure
A ready-mix plant supplied concrete for a 300 mm thick suspended two-way slab specified at 40 MPa. The approved mix contains 380 kg/m³ of portland cement, w/c = 0.44, 720 kg/m³ of fine aggregate, 1080 kg/m³ of coarse aggregate and 167 kg/m³ of water, with a water-reducing retarder. At batching, the sand moisture probe read 4.2% and the batcher deducted water on that basis. The next morning an oven-dry check showed the sand actually held 6.8%. The 45 m³ day's pour had already been placed and cured under compound when four cylinders from the day's sample broke at 31, 33, 30 and 32 MPa at 28 days, against a mix design target of 42 MPa. The slab panels concerned carry an office live load and the design moments use about 60% of the provided capacity, but the slab forms a car-park soffit with 75 mm cover and a specified maximum w/c of 0.45.
1. Quantify the batching error and assess how much of the 28-day strength shortfall it plausibly explains.
Read the worked answer
Water error. The probe under-read by 6.8 − 4.2 = 2.6 percentage points, so water already present in the sand was added again at the batcher: 0.026 × 720 = 18.7 kg/m³. Actual batch water ≈ 167 + 18.7 = 186 kg/m³, giving an effective w/c = 186/380 = 0.49 instead of 0.44, with cement content unchanged.
Strength consequence. Using Abrams' relationship with a notional strength of about 45 MPa at w/c 0.44, a shift to 0.49 typically costs 8–12% of strength, i.e. roughly 4–6 MPa. The site rule of thumb (each extra 10 kg/m³ of water costs 2–3 MPa) gives the same 4–6 MPa. From a 42 MPa mix target that predicts about 36–38 MPa.
Conclusion. The observed values of 30–33 MPa are below even that prediction. The water error is the dominant mechanism but does not account for the whole gap, so batch-to-batch variation, cement strength, and cylinder handling remain open. That is precisely why the next step must be physical evidence from the slab, not further arithmetic on the mix.
2. The contractor argues the cylinders are unrepresentative and the slab should be accepted as placed. Apply the acceptance criteria and state the next step.
Read the worked answer
Acceptance criteria (ACI 318, f'c = 40 MPa). A strength test is the average of two cylinders. Every average of three consecutive tests must equal or exceed f'c, and no single test may fall below f'c − 3.5 = 36.5 MPa.
Application. Test 1 = (31 + 33)/2 = 32.0 MPa; Test 2 = (30 + 32)/2 = 31.0 MPa. Both are below 36.5 MPa, and the running average of 31.5 MPa is far below 40 MPa. Two independent criteria fail: the concrete does not conform.
Weighing the contractor's argument. Cylinders sampled at discharge describe the concrete delivered into the slab, so low results are evidence about the slab unless curing or capping was defective. A procedural defect (cylinders cooked on site, bad caps) is a reason to question the numbers, not a demonstration of strength.
Next step. Core testing under ACI 318 26.12: at least three 70–100 mm cores from representative, structurally critical locations, avoiding reinforcement, moisture-conditioned and tested moist. Acceptance: mean of three ≥ 0.85 f'c = 34 MPa, and no core below 0.75 f'c = 30 MPa. If cores pass, accept; if not, the engineer performs a structural evaluation, and removal is only one possible outcome.
3. Identify the quality-control failure and specify the plant changes that would prevent recurrence.
Read the worked answer
Root cause. The batch water was controlled from a single, unverified moisture probe in one stockpile position. Sand drains and segregates, so surface probe readings can differ from the average moisture by several points, especially after rain or when the stockpile is rebuilt. No independent check (oven-dry or hot plate) was made before or during the pour, and no trend was compared against previous batches.
Corrective actions.
- Calibrate probes each shift and cross-check at least once per shift against oven-dry moisture of a sample from the batch being used.
- Fit two probes at different positions and depths and use the higher reading, or default to the wetter value until verified.
- Cover and drain the sand bay so moisture is stable, and re-check after any rain.
- Record moisture, batch water and slump for every truck so a step change is visible immediately.
- Bar jobsite water addition entirely; adjust slump with admixture under the supplier's approval.
Verification. Re-run a trial batch with the corrected moisture and confirm slump, yield and 7-day strength before resuming supply to structural work.