A novel approach for the modelling of high-pressure grinding rolls M. Torres, A. Casali * Mining Engineering Department, University of Chile, Chile article info Article history: Received 3 December 2008 Accepted 28 April 2009 Available online 22 May 2009 Keywords: Modelling Grinding Sulphide ores abstract The HPGR technology has become more attractive to the copper industry because of its high throughput capacities and its low specific energy consumptions. A HPGR model, able to give enough information based on pilot plant testing, in order to back up HPGR engineering studies, was developed. The model was based on the physical phenomena of the grinding operation. The model parameters were fitted with pilot scale test results, corresponding to a Chilean copper ore, classified in two lithologies (andesitic and porphyrytic ores). Some sets of data were not used in the fitting stage, to test the predictive capability of the model. The pilot scale tests were performed at the facilities of two HPGR manufacturers, changing operating pressure and rolls peripheral velocity (only one of the manufacturers). The simulated specific energy consumptions and particle size distributions, compared with the experimental data, were consid- ered good enough. The model was able to predict adequately throughput capacity, specific energy con- sumption and particle size distributions of the edge, centre and total products. Ó 2009 Elsevier Ltd. All rights reserved. 1. Introduction High-pressure grinding rolls (HPGR) technology has struggled for acceptance into the hard-rock mining sector, has had many ad- vances in this sector, but is still regarded as an ‘‘immature” tech- nology (Bearman, 2006). It has been applied to progressively harder, tougher and more abrasive materials, generally success- fully, but not without some problems (Morley, 2006). The HPGR consists of two counter-rotating rolls mounted in heavy-duty frictionless bearings, enclosed in a strong frame. Pres- sure is applied to one of the rolls that can move linearly by means of a hydro-pneumatic spring system, while the other roll is held in a fixed position in the frame (Klymowsky et al., 2002). The pressure exerted by the hydraulic system on the floating roll, that allows horizontal movement of the moving roll, largely determines com- minution performance. Typically, operating pressures are in the range of 5–10 MPa, but can be as high as 18 MPa. For the largest machines, this translates to forces of up to 25,000 kN (Morley, 2006). The rolls are driven by separate motors and can be operated at fixed or variable speed (Klymowsky et al., 2002). In most mineral applications, the roll surfaces are protected by implanting tungsten carbide studs that help to form an autogenous wear layer on the rolls and improve the drawing of the material into the rolls (Kly- mowsky et al., 2002). Roll diameters of industrial and semi-industrial units vary from 0.8 to 2.8 m. Capacities range from 50 to up to 3000 t/h. Energy consumption is between 1 and 3 kWh/t (Klymowsky et al., 2002). There are currently three recognized manufactured of HPGR ma- chines, namely Polysius, KHD Humboldt Wedag and Köppern, all based in Germany (Morley, 2006). At present, the three HPGR producers will all give guarantees of throughput and useful life of their equipment, as long as they com- plete sufficient representative test work (Danilkewich and Hunter, 2006). The test work will require obtaining ore representative sam- ples and sending around 1000–1500 kg samples to the manufac- turers and they will run HPGR amenability tests. After this phase is completed, pilot or semi-industrial testing with additional sam- ples will be required. HPGR suppliers stipulate that scale-up of pi- lot units should be done with caution. The main objectives of material testing are to determine: the ore suitability to HPGR grinding, the parameters required for sizing (specific throughput and specific grinding force), the achievable product size distribu- tion and the abrasiveness of the ore (Klymowsky et al., 2002). In terms of energy consumption, the traditional Bond theory to esti- mate the energy requirements can not be used because it grossly underestimates the actual grinding energy of the HPGR (van Dru- nick and Smit, 2006). Several tests have been developed in order to quantify the behaviour of different ores in the various crushing and grinding applications (Bond work index, JK Drop weight test, SAG power in- dex, etc.). However, none of these tests can be applied to high-pres- sure grinding (Patzelt et al., 2006). Accordingly, the only remaining alternative is the use of pilot or semi-industrial testing data. The properties of an ore have a far greater impact on achievable fines production than the grinding force (Patzelt et al., 2006). The prod- uct fineness is controlled by the grinding force applied to the mate- rial bed between the rolls, causing micro-cracks and breakage of 0892-6875/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.mineng.2009.04.011 * Corresponding author. Tel.: +56 2 9784477. E-mail address: dirdimin@ing.uchile.cl (A. Casali). Minerals Engineering 22 (2009) 1137–1146 Contents lists available at ScienceDirect Minerals Engineering journal homepage: www.elsevier.com/locate/mineng