HPGRs for hard-rock applications in Nevada

The next milestone in HPGR development came in April 2003, when Newmont Mining Corporation began a 3-month trial of a pilot-sized HPGR. Polysius designed a new roll surface specifically for the grinding of hard and abrasive copper and gold ores. The new roll surface consisted of a replaceable shrink-fitted tyre, armed with a new design of tungsten carbide studs, and a new edge-protection system intended to eliminate repair welding of the roll edges. The rolls were also provided with cheek plates to contain the material within the gap.

The roll dimensions were 950 mm diameter u350 mm width. Each roll was driven by a 160-kW motor and was operated at a speed of 21 rpm. The capacity of the unit was about 80 tph. The unit was run in closed circuit with an 8-mm square-mesh screen, and was protected from tramp metal by an overhead magnet and a metal detector on the main conveyor belt.

The machine was operated 24 hours a day, 7 days a week, for a period of 87 days with no mechanical downtimes. The initial operational availability of the unit was 89% due to a programming glitch, which occurred after startup, after which a 92% availability was achieved.

The Lone Tree trial was a true milestone, as no other pilot unit previously had been operated continuously on a 24/7 basis, and no mechanical or welding repairs had been required. The operators and maintenance staff were encouraged by the operation of the HPGR.

No stud failures occurred during the more than 1,600 operating hours of the demonstration trial. In a commercial-scale application on a similar hard, abrasive ore, an HPGR would have achieved more than 3,000 hours of service and run considerably longer on a less competent and less abrasive ore.

Relationships between wear and particle-size distribution were obtained that will improve future understanding of wear and wear life, benefiting the industry as a whole. The Polysius ATWAL laboratory abrasion test accurately predicted the wear rate in the trial. Individual tests, conducted over the course of 1 year on several representative samples obtained prior, during, and after the trial, were found to be reproducible within 10% of each other, validating the method used for determining and predicting wear in larger units.