Benchmarking: Collahuasi ball mill model

Benchmarking: Ball Mill Power Draw - Collahuasi

Villouta, R.M., Collahuasi: After years of operation, Proceedings of the International Autogenous and Semiautogenous Grinding Technology conference (SAG 2001), Vancouver, Canada, Pages I-31 – I-42

• Ball mill dimensions 22 ft diameter by 36 ft (35.5 ft effective grinding length)
• shell-supported, flat ended mills
• twin-pinion drive
• Synchronous fixed speed motor size 4850 kW (with pinion)
• Mill speed originally 74% of critical, later changed to 78% of critical.
• Ball charge 35% v/v (retaining ring)
• DCS power draw originally 8900 kW, rose to 9200 kW

Assumptions:

• ore density 2.6 t/m³
• liner thickness 6 inches
• slurry 70% solids
• motor efficiency (efficiency to DCS value) 0.942 (author's notes from a site visit)

	DCS Power, kW	Speed, % of critical	Mill filling	Morrell Model, kW	Nordberg model, kW (no density correction)	Nordberg model, kW (density correction)
original	8900	74%	35% v/v	9044 (+1.6%)	[deduct 5%] 8183 (-8.8%)	[deduct 5%] 8717 (-2.1%)
later	9200	78%	35% v/v	9655 (+4.9%)	[deduct 5%] 8890 (-3.5%)	[deduct 5%] 9471 (+2.9%)

All power values relative to DCS.

Discussion

The Nordberg models aren't really set up for flat-ended mills; the empirical basis of Nordberg models are cone-ended mills. Literature suggests a typical 5% factor be applied to convert a flat-ended mill (shell-supported) to a cone (trunnion-supported). The values in the table deduct 5% from the raw values shown in the Nordberg model.

As this is a somewhat atypical mill configuration, the Overflow ball mill using full Morrell C-model is recommended as the cone angle is a parameter of the equations. It is possible to get the Nordberg model to mimic the 5% reduction shown above (by modifying the mechanical efficiency factor), but not recommended.