A fast numerical method for evaluating the performance of float-sink coal-cleaning equipment

Abstract

A method is presented which seeks to overcome a number of disadvantages that are inherent in the conventional practice of plotting the Tromp curve and analysing it for its various measures, e.g., specific gravity of separation, probable error and error area. A sufficient condition is established for extracting the appropriate criteria of the sharpness of classification, which are independent of coal-feed distribution in specific gravity. Partition area, a modified areal measure, similar to the conventional error area, of the equipment's deviation from ideal classification behavior is derived from the theoretical partition function. A purely numerical procedure is developed for locating the specific gravity of separation and the partition area. Unlike the Tromp-curve-based conventional procedure, the proposed histogram-based method is fast, more convenient and objective for analysing the coal-cleaning plant data.