Keywords: computational fluid dynamics, cyclonic separator, measurement, empirical form, simulation.

It is very compicated to describe the air flow in cyclonic separators, since they are closed objects: so the working of it cannot be seen. The voricity is high and can be described as turbulent, plus there are extra component within the air, the dust, which are to be separated. Computational fluid dynamics is a good tool to investigate the flow in cyclonic separators, but there are many challenges. First the modelling and meshing process gives us some barriers to overcome, because we already know that the most important and questionable section is the inside wall of the cyclon, since the separation takes place there, so the boundary layer flows have to be modelled very carefully. For this reason a boundary mesh (prism layer) is also necessary, which greatly increases the number of elements. The modelling of both phases and the turbulence needs special multi-phase and turbulence models to be used.

The simulation has been calculated on a grid, made from three computers, the volume is didvided into three domains. A RANS calculation with k-epsilon turbulence model and mixture multi-phase model were used. The validation was made by measurements, while the whole project is based on the cyclonic separator located at our university. It is possible also to validate the simulation by traditional calculations. The method was developed by Barth and Muschelknautz, and mainly contains empirical formula, which contains basically 10-14% error.

Therefore, the cyclone pressure drop is around 1898 pascal, greater by 311 pascal than the measured value and by 992 pascal as simulation calculated 906 pascal. The difference has arisen as a result of the modelling of cyclonic separator where the vortex tube widens upwards like a diffuser, which is referred to as a pressure drop reduction factor in the literature.

The simulation is partially successful while more problems became apparent. The predicted pressure drop is inaccurately predicted, the test-powder concentration is higher in the outlet than in the inlet, which is attributable to the RANS model, but the simulation provided the correct flow field from the air's side, in comparison with the measurements.

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £40 +P&P)