Precise predictions of particle motion behavior after injection into a fluid flow are essential for process design and optimization in industrial applications. The goal of this work is to present a numerically inexpensive CFD model to predict lateral particle distribution after injection into a vertical cross flow. To this end, a specially designed and application-oriented flow channel, a so-called particle cross sifter, was built, which was experimentally and numerically analyzed. In the experiments a type of boiler slag powder, consisting of particles with various sizes and shapes, was injected into the cross sifter. For the numerical simulations different turbulence modeling approaches, such as RANS and LES, in combination with the DPM were applied. The results were compared to experimental data. It is highlighted, that only the LES approach in combination with the DPM is able to predict the lateral particle distribution with sufficient accuracy. (C) 2020 Elsevier B.V. All rights reserved.