Cyclone optimization including particle clustering

Abstract

In this work, a new family of geometries of reverse-flow cyclones was obtained through numerical optimization, using a stochastic random search global optimizer coupled with the PACyc model. The objective was to optimize the geometry of a reverse-flow cyclone taking into account inter-particle agglomeration (clustering), since this phenomenon usually occurs to some degree in industrial cyclone operation, increasing the collection of fine particles.Experimental results for three kinds of particles and particle size distributions are shown using a pilot-scale unit. An industrial implementation of the new optimized cyclone is described and the results concerning the performance of the system are shown and compared with predictions from the PACyc model.The results show a highly improved global efficiency when compared to that of a cyclone geometry obtained by a similar optimization methodology while neglecting the agglomeration/clustering effect. This opens the possibility of using reverse-flow cyclones to capture very fine particles, complying with strict emission limits, such as those from biomass boiler exhausts.