Models which have been presented in the literature consider only selected aspects of dense medium cyclone operation, are empirical in nature and generally relate only to severely limited ranges of operating parameters. This paper reports progress towards development of a more comprehensive and robust model. The data base consists of partitioning parameters developed using density tracers and conventional coal sampling and analysis techniques, together with recordings of parameters such as cyclone dimensions, feed pressure, medium size distribution and product media densities.. These data were drawn from 15 Australian coal preparation plants._x000D_
To elucidate the effects of important operating parameters on medium behaviour and on the partitioning of large particles a series of pilot scale tests was also conducted. The present model consists of eight sub-models which relate to particular aspects of dense medium cyclone operation such as medium flowrates and densities, or size- by-size separation densities and Eps. The model structure follows a logical progression which clearly demonstrates the importance of understanding medium behaviour before attempting to predict partitioning performance. The model also attempts to define the conditions under which yield losses due to unstable "surging" discharge may occur.