Segregation is a very common phenomenon in industrial processes wherever granular materials are handled. It can occur at different stages in an industrial process and it is usually undesired. To support the industrial design, it is necessary to understand the mechanisms of segregation and to develop a tool which is able to predict the propensity of segregation depending on product and process properties. The underlying mechanisms of segregation during filling and discharge of silos are discussed in detail. To study the effects for different configurations of silos (flat bottom vs. inclined hopper; with and without belt conveyor) a bimodal mixture of differently sized particles is employed. A useful tool to support the analysis of segregation processes is the discrete element method (DEM). A calibrated and validated DEM simulation has been used to simulate a pilot scale silo with a belt conveyor at the outlet. In this context, the influence of some common procedures in DEM, which are used to reduce the computational effort (way of filling, upscaling, periodic boundary condition) as well as their effect on the measured segregation is presented. Moreover, a continuum approach is presented which fits the segregation effects using a simplified model to experimental data. This approach is useful in a first stage to predict the segregation behaviour based on a convective/diffusive model. (C) 2015 Elsevier B.V. All rights reserved.