Grinding tests with two pure ores, quartz and marble, were conducted in mono and polysized samples in order to develop a mathematical model of the high pressure roll mill (HPR) size reduction performance for use in simulation and design studies. The effects of feed size distribution, rolls grinding force, rolls speed, initial gap and moisture content in the feed were investigated.
The problem of sizing the agglomerated HPR product was realised and properly addressed by assessing the true size distribution using ultrasonic energy. A simplified tumbling test method was further developed for standardize the sizing of all tests products.
The experimental data was used for testing the self-similarity approach (Fursteneau et al, 1993) for prediction of the HPR product size.
A new model was developed based on the Whiten general crusher model structure, using breakage function data from HPR lab tests and bed breakage tests conducted in a piston and die device. The model was conceived in three parts: the single particle crusher, where it is assumed that particles bigger than a certain gap are subjected to breakage in the single particle mode; the edge effect crusher, which receives a fraction of the product of the single particle crusher (material that is coming from the edge of the rolls) and perform further breakage to the particles bigger than the measured working gap, also under the single particle breakage mode and, finally, the true HPR crusher, which receives the fraction of the single particle breakage crusher product that is not coming from the edges of the rolls. In this crusher, breakage occurs in a highly compressed particle bed, and the breakage function data was obtained from tests undertaken in a piston and die device.
The HPR model was used to predict the product size distribution of the model ores, quartz and marble, and ultimately, it was validated using data from a typical Australian gold ore (Boddington material).
A power model was also developed using the Andersen's (1988) pendulum power relationship to predict the power draw by the rolls during each test. This power determination coupled with the working gap calculation undertaken by another developed routine (the spring-program) may be used for scale up of results to help in industrial scale machine design.