Aggregate Quality in a Deccan Basalt Quarry: Rejecting the 'Average'
A contractor opens a new bench in a Deccan basalt flow near Bagalkot to supply concrete aggregate for a bridge. The flow exposes three zones over 40 m: (1) a lower massive, fine-grained, closely jointed basalt; (2) a middle vesicular basalt in which vesicles up to 8 mm are lined or filled with white zeolite, green clay and, in places, chalcedony; (3) a top red bole and clay-rich weathered basalt. The client's specification requires Los Angeles abrasion ≤ 30%, water absorption ≤ 1.0%, specific gravity ≥ 2.60, and no potentially alkali-silica reactive constituent. Preliminary tests give: zone 1 - LAA 21%, absorption 0.45%, SG 2.94; zone 2 - LAA 33%, absorption 2.6%, SG 2.53; zone 3 - LAA 41%, absorption 4.1%, SG 2.31. The quarry manager wants to blend all three 'because the average will be fine', and points out that the vesicular rock is the easiest to rip.
1. Test the 'average will be fine' argument numerically and against specification practice.
Read the worked answer
Simple three-way average: LAA = (21 + 33 + 41)/3 = 31.7% - already fails ≤ 30% Absorption = (0.45 + 2.6 + 4.1)/3 = 2.38% - fails ≤ 1.0% SG = (2.94 + 2.53 + 2.31)/3 = 2.59 - fails ≥ 2.60
So even the manager's own arithmetic fails all three criteria; the averaging argument is not merely poor practice, it is arithmetically wrong.
Why averaging is wrong in principle:
- The specification applies to the aggregate actually placed, and a blend has a distribution, not an average. One weak particle in the mix breaks; the surrounding particles do not average it out.
- Weak vesicular and clay-rich particles crush under compaction and traffic, generating fines, losing interlock and producing dust - a durability failure, not a strength failure.
- Density sorting during stockpiling and transport concentrates the light vesicular lumps in the coarse fractions, so the placed material can be worse than the blend tested.
Zone 1 alone meets every criterion; zones 2 and 3 fail individually and are not made acceptable by dilution with good rock.
2. Explain petrologically why zone 2 fails, and what the zone-3 bole tells you about the flow.
Read the worked answer
Zone 2 is the amygdaloidal top of the flow.
- Vesicles form by degassing as the lava surface cools and pressure drops; the upper part of a flow is therefore always the most vesicular and the most porous.
- The vesicles are then filled by secondary minerals - zeolite, calcite, smectite clay and chalcedony - precipitated from late fluids. These minerals are soft, cleavable and alterable; they lower specific gravity (2.53 against 2.94) and raise absorption (2.6% against 0.45%).
- The chalcedony is cryptocrystalline silica, which reacts with the alkaline pore solution of concrete - an alkali–silica reaction risk that alone disqualifies the rock from bridge concrete, regardless of its abrasion number.
- Weathering along vesicle walls and joints has converted much of the interstitial glass to clay, so the rock is not merely porous, it is partly decomposed.
The red bole in zone 3 is a palaeosol - the weathered top of this flow, buried by the next lava. It marks a flow boundary, so the section is one flow: massive colonnade below, vesicular top, bole above. The next flow's massive base lies above the bole and may be good rock - worth mapping before abandoning the bench.
3. Give the quarry recommendation and the sampling rule that would prevent this argument recurring.
Read the worked answer
Recommendation:
- Quarry zone 1 only for structural concrete. It meets LAA, absorption, SG and has no reactive silica.
- Strip and reject zone 3 as overburden; it is clay-rich weathered rock, not aggregate.
- Stockpile zone 2 only as non-structural fill, and only after a petrographic examination confirms no chalcedony, plus an ASR mortar-bar test. It must never enter bridge concrete.
- Rippability is not a quality argument. Cheap excavation of a bad rock is paid for at the batching plant: rejected concrete, retesting, delay and liability. Ripping cost is a few tens of rupees per tonne; a failed pour is orders of magnitude more.
Sampling rule that prevents the argument:
- Delineate the flow units on the bench face and map them; quarry by flow unit, not by quarry.
- Sample each bench, over its full height (channel or core) at 5 m intervals, not grab samples from the easy rock.
- Record the geological log with the test certificate so a future blend proposal can be checked against the flow stratigraphy rather than judged on a single number.
Test the bole's overlying flow base before extending the pit - the good rock may be nearer than expected.