Numerical simulation of droplet impingement and flow around a cylinder using RANS and LES models

•Droplet impingement on structures in arctic climates leads to icing hazard.•Numerical models of droplet flow are used to estimate icing.•Simple flow models predict well impingement of droplets on the cylinder front.•Most RANS models predict droplet flow in the wake region different from LES.•The k–ε model with the enhanced wall treatment predicted similar result to LES.

The paper is devoted to the numerical modeling of droplet impingement on a cylinder and droplet transport behind it at a Reynolds number of 1.6×104. The objectives are to investigate the ability of different turbulence models to predict the collision efficiency and the droplet distribution in the wake in order to find the computationally most efficient way to achieve reliable results. The solutions of potential theory and four different RANS models are compared using the one obtained with Large Eddy Simulation as reference. The results show that the collision efficiency of the cylinder is predicted well by both potential theory and all considered RANS models for the given Reynolds number and droplet diameters of 13, 17, and 45 μm. However, only the RNG k–ε model with enhanced wall treatment results in the droplet flow in the wake behind the cylinder being similar to the results obtained by LES.