Fragmentation Prediction · Grain

Pattern and charge

The geometry Kuz-Ram reads. Everything downstream moves when these move.

Burden m

Spacing m

Bench height m

Hole diameter mm

Charge length m Explosive column only. Stemming is not charge.

Charge per hole kg

Relative weight strength ANFO=100 Emulsions 100–115, ANFO 100, heavy ANFO 105–115.

Rock density t/m³

Drilling accuracy W m Standard deviation of hole position at the toe. Cunningham's (1 − W/B) term.

Holes on the shot count

Rock factor A

Where all the uncertainty lives. Built from the Lilly components, or entered from a calibration blast.

RMD — rock mass description

JPS — joint plane spacing

JPA — joint plane angle

UCS MPa Hardness HF = UCS ÷ 5 when Young's modulus is 50 GPa or more.

Young's modulus GPa 0 = not known. Below 50 GPa, Lilly uses HF = E ÷ 3 instead.

Override A from a calibration blast Back-calculated from a measured x50 on this bench, in this rock.

Predicted passing curve

Rosin–Rammler and Swebrec on the same x50 and n. They agree at the middle and separate in the fines. A ±20% band Swebrec

Design comparison

Nobody wants one prediction. They want to know whether tightening the pattern is worth it.

Sensitivity to powder factor

Burden and spacing scaled together at constant charge per hole — the trade a pit actually makes.

Oversize and fines

Tonnes above and below your own cuts, and the boulder count that follows from them.

Crusher opening m Anything above this is a boulder that has to be handled twice.

Fines cut m

Fragment shape factor Volume = factor × d³. 0.52 is a sphere; angular rock sits near 0.5.

Secondary breakage $/boulder Your number. Leave at 0 and no dollar value is reported.

Crushing + haulage $/t

Drill and blast $/hole Without this, a tighter pattern always looks cheaper.

Prediction

—

The pattern, the charge and the rock factor above produce the prediction here: x50, the uniformity index n, P20 to P90, and the percentage above your oversize cut and below your fines cut.

Every size is reported against the rock factor A ±20% band as well, because that band is wider than most of the design changes anyone is choosing between.

Lilly components

A = 0.06 × Σ

What A does to the answer

A versus B

no reference stored

Set a design, press Store design A, then change the pattern. Both curves are drawn together and the differences are tabulated.

The comparison is the point. A single predicted P80 carries an uncertainty far larger than most of the changes an engineer is choosing between; the difference between two predictions made with the same rock factor is far better conditioned than either number alone.

What this does and does not do

Stated limits

attached to every result