A CFD study on the design, optimization, and scaling of cyclone separators for steam-droplet separation

Abstract

For secondary steam-droplet separation in drum-type heat recovery steam generators, a multi-inlet cyclone separator is used, because of its robustness, compared to the state-of-the-art demister pads. A corresponding challenge is to limit the pressure drop while maintaining high mass collection efficiency.

A CFD approach applying the Reynolds Stress Turbulence Model and Eulerian-Lagrangian multiphase model is proposed and validated for predicting the flow field and performance. This approach was used in a multi-objective, design-of-experiments-type optimization study, applying the response surface methodology (RSM) for surrogate modelling and obtaining the Pareto Front.

For fixed design conditions, a geometry with 37% lower pressure drop was obtained, while maintaining the required separation performance. For scaling and dimensioning to different design conditions, two scaling methods are proposed. In addition, RSM-based models are obtained, which are able to predict the (dimensionless) pressure drop with errors below 5%, grade efficiency curve, and mass collection efficiency for distinct design conditions.