Heavy Media Selection Functions - Circuit Analysis

Current coal cleaning technology uses heavy media (HM) for the recovery of coal in particles larger than 1/16 of an inch. Through the use of float sink measurements of the feed and products to commercial HM unit operations, a selection function, sometimes called Tromp curve in the coal industry, can be computed. Since the selection function is a plot f the probability of a particle appearing in the float versus the particle density, the slope of the curve will be sharper at the point where the particle has a 50% chance f appearing in the float. Thus, the separation will be better. Fine-tuning a unit operation or a circuit of unit operations requires that the selection function be accurately known. By current practice, it is difficult to obtain an accurate measure of the slope because, when the selection function is plotted versus density, the data forms an S-shaped curve which lacks data points near the 505 point of the curve. This lack of sensitivity in measuring the slope means that small recovery improvements cannot be detected; thus, only large improvements are measurable. Since most engineering improvements are small when individually considered, but large when aggregated ever many small improvements, this lack of measurement sensitivity rules out the slew march toward improved efficiency. This paper presents a method of analyzing heavy media selection functions resulting in a far more effective measure of the selective function slope. Furthermore, this improved method is applied to a number of industrial coal cleaning plants for both heavy media and heavy media cyclones. Using the work of Hudy (1968), and Deurbrouck and Hudy (1972), selection functions for individual unit operations in each plant are calculated as a function f particle size as well as density.