Sampling a Stockpile for a Gradation Acceptance Test
A technician must sample a 4,000-tonne crushed-stone stockpile for gradation acceptance. The stockpile is a conical pile about 12 m high; the outer 0.5 m has visibly segregated, with coarse material at the base and fines near the top. The specification requires a sample of at least 25 kg for the nominal maximum size, reduced to test size. The technician plans to take one shovel sample from the base of the pile, near the conveyor discharge point, because that is where the coarse material is and 'that is what the specification is about.' The previous acceptance sample from this stockpile passed. The producer is waiting to ship.
1. Critique the sampling plan and describe a defensible alternative.
Read the worked answer
The plan is not defensible. A single shovel sample from the segregated base of a 12 m conical pile is biased toward coarse material and is not representative of the stockpile. The base is exactly where segregation concentrates coarse particles; sampling there guarantees a gradation skewed coarse and could produce a false failure (or, if the fines were sampled, a false pass).
A defensible plan: (a) treat the stockpile as a lot and take multiple increments from different locations and depths - at minimum, from several points around the pile and from the interior, not just the surface; (b) avoid the segregated outer 0.5 m where possible, or sample it deliberately and document that; (c) combine increments into a composite of at least 25 kg for the nominal maximum size, then reduce by quartering or splitting to test size; (d) record locations, depths, and number of increments. If safe access to the interior is impossible, sample from the conveyor or from a stockpile being built, which is more representative than a finished cone.
The prior passing result is not evidence for this sample; it is a different sample from a pile that has since segregated further.
2. The composite sample is taken and the 4.75 mm passing is 8%, within the 0–10% limit. The producer asks whether the result can be used for the whole 4,000 tonnes. What is the correct answer?
Read the worked answer
A single composite sample characterizes the material sampled, not automatically the entire 4,000-tonne lot. Whether it can be applied to the lot depends on the sampling plan: if the increments were drawn from across the lot in a documented, random or stratified pattern, the composite can support a lot-level statement with the associated sampling uncertainty. If the increments were clustered, the result applies only to what was sampled.
Also, a result of 8% against a 10% limit carries little margin. Combined with the known segregation in the pile, the honest answer is that the result is consistent with acceptance but does not demonstrate uniformity across the lot. The correct response is to state the scope of the result, recommend additional sampling from other parts of the pile or from the conveyor during shipment, and not to certify the whole lot on one composite. The producer's schedule pressure is not a technical basis for extending the result.