Benchmarking: Yanacocha Single-stage SAG Mill

Benchmarking: SAG Mill Power Draw - Yanacocha

• Burger, B., Vargas, L., Arevalo, H., Vicuna, S., Sidel, J., Valery, W., Jankovic, A., Valle, R. and Nozawa, E., Yanacocha Gold Single Stage SAG Mill Design, Operation, and Optimization, Paper № 127 in Proceedings of the International Conference on Autogenous Grinding, Semiautogenous Grinding and High Pressure Grinding Roll Technology (SAG 2011) held September 25 – 28, 2011 in Vancouver, B.C., Canada.

Mill geometry and power-draw modelling

Single-stage SAG mill geometry given as:

• mill nominal diameter (inside shell): 32 ft
• mill effective grinding length: 32 ft
• liner effective thickness: 175 mm (7 inches), eyeballed from new liner diagram

Ore density is 2.52 t/m³

Not stated in the paper is that the 16.5 MW motor is gearless with DCS indication assumed to be motor output power.

	Shell Power, kW	Mill Speed, RPM (%crit)	Ball charge, %v/v	Filling, %v/v	Feed %solids	Simplified Morrell model, kW	Loveday/Barratt, kW	Austin model, kW cylinder + 5%
First Survey	12,286	8.9 (64.5†)	16.5	17.9	73	11,587 (-5.7% difference)	15,005 (22.1% difference)	13,020 (6.0% difference)
Second Survey	13,992	8.7 (63.1†)	19.1	22.9	80	13,067 (-6.6% difference)	15,858 (13.3% difference)	14,496 (3.6% difference)

^† The %critical speed values given in the text do not appear to account for the liner. The RPM values are assumed to be correct and the %critical speed values have been back-calculated.

Simplified Morrell SAG mill model was run, default k=1.26 (motor input basis) used. Austin model also used default fitting factors K=10.6 and A=1.03.

Specific Energy Consumption

The Burger et al, 2011 paper gives results of two mill circuit surveys, the first from May 12, 2010 and the second from June 9, 2010. It reports that the mill was not operating efficiently during the first survey — "Grate pegging was significantly more severe during the first survey, and is considered to be a major contributing factor to the difference in performance for the two surveys." Also "Significant slurry pooling was observed in the Yanacocha SAG mill during both of the surveys."

Grindability test results

Grindability measurements of survey samples are are given as:

• Bond ball mill work index is reported as:
  • 17.53 kWh/t (106 µm closing screen, assume P₈₀=75 µm),
  • 16.52 kWh/t (212 µm closing screen, assume P₈₀=155 µm).
• Point Load Index (PLI) average Is₅₀ is 2.62 MPa.
  • Assume an impact crushing work index of approximately 8.0 (metric).
• Bond abrasion index (Ai): 0.528
• SMC test results: DWi 3.47 kWh/m³, Mia: 12.3 kWh/t, Mih, 8.0; Mic 4.1 kWh/t; A: 82.8; b: 0.88; density: 2.52 t/m³; t_a: 0.75
  • Assume a rod mill work index of approximately 15.2 (metric).

Some manipulation of the ball mill work index results is required to correct the work index result to plant target P₈₀. This is done by fitting the two tests to an equation of the form Wi(x) = a x^b where x is the test P₈₀ and a and b are fitting parameters. Fitting the equation gives a=25.65 and b=-0.09. These parameters are entered into the Bond/Barratt circuit configuration and the resulting ball mill work index used for modelling is 16.4 metric units (almost identical to the 212 µm closing mesh sample).

Wi(x) = 25.65 × (P80, µm)^-0.09

Model predictions

The reported throughputs and specific energy consumption (E_SAG)

• Survey 1: 620 tonnes/hour; 19.8 kWh/t
• Survey 2: 779 tonnes/hour; 18.0 kWh/t

The Bond/Barratt single-stage SAG model results in the following:

• Survey 1: F₈₀=79.1 mm; P₈₀=152 µm; E_asag= 16.4 kWh/t (20% low)
• Survey 2: F₈₀=72.4 mm; P₈₀=154 µm; E_asag= 16.3 kWh/t (10% low)

The Morrell Mi single-stage SAG model using the assumed ball mill work index F₈₀ and grams per revolution from the 212 µm closed test results in the following:

• Survey 1: F₈₀=79.1 mm; P₈₀=152 µm; E_asag= 12.7 kWh/t (56% low)
• Survey 2: F₈₀=72.4 mm; P₈₀=154 µm; E_asag= 12.55 kWh/t (43% low)

Checking the Mi calculation

Both of these Mi results are significantly less than the survey. It is worth manually checking to see if the SAGMILLING.COM software has made an error in these calculations.

First is to check the Mib value. The ball mill work index test closest to the target P₈₀ size is the 212 µm closing mesh sample. Given:

• closing mesh size: 212 µm
• work index: 16.52 metric units

We can add a couple of assumptions and work out the Mib value:

• test feed F₈₀ is 2300 µm (a typical feed size)
• test product P₈₀ is 155 µm (a typical product size for a 212 µm screen closing the test)
• test grindability is 1.58 g/rev (by calculation)
• Mib = 18.18 / ( 212^0.295 × 1.58 × (155^-0.295 - 2300^-0.297 )) = 18.85 kWh/t.

The Mia value is reported directly from the SMC Test™ as 12.3 kWh/t. For Survey 1:

• E_coarse=12.3 × 4 × (750^-0.296 - 79100^-0.374) = 6.22 kWh/t
• E_fine = 18.85 × 4 × (152^-0.295 - 750^-0.296) = 6.47 kWh/t
• E_peb crush = 0 kWh/t
• Etotal = E_coarse + E_fine + E_peb = 6.22 + 6.47 + 0 = 12.69 kWh/t

Conclusion: the software is giving the same answer as the Mi calculate done by hand.