Otimização da separação sólido-líquido em hidrociclones mediante modificações geométricas

Abstract

Because of advantages such as low cost, simple structure, large capacity of suspension processing and small dimensions, hydrocyclones are widely used for separating dispersed phase (liquid or solids) from continuous one. The shape and size of a hydrocyclone has a decisive effect on the internal flow structure of the continuous phase, and therefore, the separation or classification of the dispersed phase. Therefore, several geometric changes have been suggested to improve the particle separation efficiency and reduce energy costs of hydrocyclones. In a work of hydrocyclones optimization, Vieira (2006) studied the influence of the main geometrical variables on the performance of the separators, from what he found a hydrocyclone with good separation efficiency and low energy cost, the so-called HC11. In the present work, new geometrical changes have been incorporated to the hydrocyclone HC11 and the impacts of those changes have been evaluated experimentally and by CFD techniques. The proposed geometrical changes were: the modification of the wall thickness of vortex finder; the use of a mantle type vortex finder; the use of a ramped roof feed inlet and the use of a rotating feed. The results indicated that all the proposed modifications have changed the performance of the separator and that it is possible improve the standard HC11. Moreover, based on the data of Vieira (2006), three new geometrical configurations of hydrocyclones have been found through the use of response surface technique combined with the Differential Evolution algorithm. The results obtained through the optimization techniques have been validated by experimental data. The optimized configurations of hydrocyclones found were: (i) hydrocyclone HCOT1, with high separation efficiency (η=85.5%); (ii) hydrocyclone HCOT2, with low underflow-to-throughput ratio (RL=9.37%) and (iii) hydrocyclone HCOT3, with low Euler number (Eu=788) and low underflow-to-throughput ratio (RL=5.08%).