The effective filtration of inlet air from dust particles is an important requirement for the
efficient engine life cycle for internal combustion engines. Axial cyclone separators are used
as primary filtration components in many such engines, and it forms a primary filter element
component in armoured combat vehicles, especially for desert applications. In this type of
cyclone separators, a swirling motion is imparted to the incoming flow with dust by means of
a swirl generator and particle separation is achieved by centrifugal separation. In the current

study, the separation of incoming dust particles is numerically investigated using an Eulerian-
Lagrangian approach. The Reynolds Averaged Navier-Stokes(RANS) equations are solved for

the fluid phase with RNG K-epsilon model turbulence closure model, and Discrete phase
modelling (DPM) is done to solve the particle motion using the open source Computational
Fluid Dynamics (CFD) tool OpenFOAM. The validation of the model is done by comparison
with benchmark cases and experimental data.
The four significant geometric parameters in an axial swirl tube cyclone separator with respect

to improving the overall performance are identified to be the blade angle, blade length, blade-
tube distance and the number of blades. The impact of these parameters on the two important

output parameters of a cyclone separator, the pressure drop and filtration efficiency, is studied
through numerical analysis. A one factor analysis is performed to understand the individual
contributions of the parameters and fi is shown that the blade length is found to be the most
sensitive parameter, followed by the blade angle, number of blades, and blade tube distance
respectively.
A Multiobjective optimisation is done using the Design of Experiments (DoE) approach. After
performing Taguchi and ANOVA analysis, a second order regression model was developed to
predict pressure drop and filtration efficiency and this model and is used as a surrogate model
for optimization. With the objective of obtaining minimum pressure drop and maximum
filtration efficiency, a set of 18 optimum configurations was obtained as a Pareto front using
Genetic algorithm codes, and the results are validated with CFD simulation data. The Pareto
front points are found to have a better overall performance compared to the reference model,
and provides the designer with multiple choices for a configuration of improved performance.
Parallely, a preliminary heat transfer analysis of the flow through the filter is done with different
inlet fluid temperatures for small temperature differences, the effect on filtration efficiency is
found to be less significant. Finally, a clustered configuration analysis with upto nine different
filter elements stacked together is also done to replicate the actual filter arrangement.

Keywords: Axial cyclone separator, Computational Fluid Dynamics, Discrete phase
modelling, Multi-objective optimization, Pareto front