Experimental and numerical study of the effect of particle size distribution on hydrocyclone classification

Abstract

Classification by particle size using hydrocyclones is a fundamental step in various industries, yet the influence of material properties remains an area of limited understanding. This paper presents a comprehensive experimental and numerical investigation to illuminate the complex relationship between the feed particle size distribution (PSD) and the classification process. The feed PSD is characterised by two crucial parameters: the median size (d50f) and the distribution coefficient (σJ), derived from Johnson’s SB function. The results reveal non-monotonic variations in cut size and curve slope during physical tests where the two parameters are not separated. With one parameter held constant in numerical tests, an increase in d50f generally leads to a decrease in pressure drop but an increase in water split, while σJ exhibits an opposite trend. The interplay between the two parameters and their effects on cut size and cut sharpness is intricate, primarily determined by the characteristics of the flow field and the relative sizes of the cut size and d50f in the feed. These findings underscore the critical role of feed material properties in operating hydrocyclones for fine particle classification.