Proceedings > Papers by author > Pasternak Lars

Modelling the occurrence of bouncing in droplet collision for different liquids
Maohong Sui  1@  , Martin Sommerfeld  2@  , Lars Pasternak  2@  
1 : Otto-von-Guericke-Universität Magdeburg, Multiphase Flow Systems, Institute for Process Engineering
Zeppelinstraße 1, D-06130 Halle (Saale), Germany -  Germany
2 : Otto-von-Guericke-University Magdeburg, Multiphase Flow Systems, Institute for Process Engineering
Zeppelinstraße 1, D-06130 Halle (Saale), Germany -  Germany

Spraying systems are of great importance in a range of different technical and industrial applications. Depending
on the operational conditions and the spray structure the produced droplet size spectrum may be largely affected
by collisions between droplets. For the purpose of a numerical prediction of spraying processes by the
Euler/Lagrangian approach, the reliable models are required for predicting the collision outcome. This is mainly
done by using so-called collision maps to remark the different outcome scenarios by appropriate boundary lines
(i.e. bouncing, separation and coalescence). These boundary lines should be reasonably general including all
relevant influential effects, such as impact conditions, droplet size ratio and liquid properties. Therefore, a variety
of detailed experimental studies on the collision of higher viscous pure liquids and solution droplets were used for
developing a model for the boundary line between coalescence and stretching separation. However, the boundary
line for bouncing, mostly used in numerical studies so far was derived based on experiments with ethanol droplets
[1]. And it completely neglects viscous dissipation. Therefore, the deviations for predicting the region of bouncing
are quiet large for liquids with high viscosity or different properties. In this work, a new, more general correlation
for the bouncing boundary is derived, which however includes new assumptions and definitions. The new boundary
line is based on the studies of Estrade et al. [1] and Hu et al [2]. The original model has already included a shape
parameter 0 that actually also should be dependent of the impact parameter B which reflects the deformation during
the collision process. Moreover, an additional parameter has included to reflect the degrees of dissipation or
the energy conversion during collision. Both parameters could be linearly correlated with the impact parameter B.
The involved slope and initial values could be very well correlated with the Ohnesorge number and approximated
by third order polynomials which fitted the available experimental results. However, the slopes and initial values are
separately for pure fluids and solution droplets. The new bouncing boundary line has been developed based on
experimental result by Kuschel and Sommerfeld [3][4], Pasternak and Sommerfeld [5]. Consequently, the boundary
line for bouncing is predicted more accurately and the trends with changing liquid properties is very well captured.


Online user: 62 RSS Feed