Multi-phase modelling of hydrocyclone: validation of particle size distributions against sampling probe data

Abstract

The flow behavior in hydrocyclone is quite complex. This complexity of fluid flow in a cyclone is basically due to the fact flow in cyclone is a swirling turbulent multiphase flow. A multiphase CFD model with individual sub models for the air-core, turbulence, and particle classification was used to simulate the various hydrocyclones. The predicted velocity field from the LES, DRSM turbulence models is compared with LDA measurements of Hsieh (1988) data for 75 mm cyclone. The LES with the modified Mixture model, particles with a feed size distribution is seems to promising in predicting classification curve accurately. Further the CFD model validation is attempted to compare with Renner (1976) data which was originally measured using high-speed sampling probe at different precisely controlled positions. A CFD multi-phase case was set-up with the same 6 inch cyclone design (Renner 1976) and operating conditions. The overall classification curve predicts close to the experimental data. From the validation of CFD with the Renner (1976) data it was observed that the predicted position sample size distributions are reasonably comparable with experimental data nearby wall positions and at bottom cone section, but considerably differ at nearby forced vortex region, and also nearby tip of vortex finder wall locations. Close to the forced vortex (inner position), the discrepancy between measured and predicted size distributions are large. This may be an effect of sampling turbulence due to probing the sample close to the unstable forced vortex. Usually the forced vortex is stable with an air-core formation. The CFD model shows no air-core formation at this low operating pressure, which suggests an un-usual/un-stable forced vortex based cyclone separation in this particular experiment.