Assessment of turbulence models for single phase CFD computations of a liquid-liquid hydrocyclone using OpenFOAM

ABSTRACT

Hydrocyclones are widely used in industry and CFD has been used to compute them. Reynolds stress turbulence models (RSM), which are computationally costly and oftentimes hard to converge, are often recommended in these computations. The present work has selected a liquid-liquid separation hydrocyclone for which single-phase experimental tangential and axial velocity profiles are available. CFD has been employed to test simpler turbulence models than the RSM and results have been compared with experimental data. The turbulence models assessed in the present work were: standard k-ε, standard k-ε with a curvature correction term, RNG k-ε, realizable k-ε, k-ω, SST, a two-time-scale linear eddy viscosity model, nonlinear quadratic and cubic k-ε eddy viscosity models and the Gibson and Launder and LRR Reynolds stress models. Computations have been carried out with OpenFOAM${}^{\circledR }$ 2.2.2. Results using the Gibson and Launder turbulence model have been compared to some obtained with Ansys${}^{\circledR }$ Fluent and these were in agreement. Results have shown that all turbulence models, apart from the RSM, returned basically the same tangential velocity profiles as the standard $k-ϵ$ model. All turbulence models have failed in predicting axial velocity. Assessment of the Reynolds stresses has indicated that the internal flow field in hydrocyclones might be shear dominant and that the Reynolds shear stress component ${\tau }_{yz}^t$ is the most relevant to correctly predict tangential velocity. Geometric proportions of hydrocyclones may affect significantly the intensity of rotational and streamline curvature effects. Two-equation eddy-viscosity models are likely to be able to attend such condition, since appropriate levels of eddy viscosity are predicted at free and forced vortexes regions, however further investigation is still needed.

Keywords:

• CFD
• hydrocyclones
• turbulence models
• OpenFOAM

Acknowledgements

This study was financed in part by the Coordenação de Aperfeiçamento de Pessoal de Nível Superior – Brasil (CAPES) – Finance Code 001. The authors also acknowledge Dr. Braga for providing the computational mesh he used in his work.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by Coordenação de Aperfeiçoamento de Pessoal de Nível Superior – Brasil (CAPES) [Finance Code 001].