The ability to predict segregation of the solid phase in processes such as powder injection molding and injection molding of semi-solid materials is of special interest since such phenomenon affects the final properties and characteristics of the molded parts. In powder injection molding, for example, defects appear very often in the debinding and sintering stages but are caused by filling problems and determined by a non-uniform distribution of the solid particles within the molded part. In this paper we propose a 3D numerical solution algorithm for the simulation of particle migration in dense suspensions. The particle migration is modeled using the diffusion flux model and integrated into the NRC's 3D injection molding software. The solution algorithm is validated by solving flow problems for which experimental and numerical data are available: circular Couette flow, piston driven flow and sudden contraction-expansion flow. Since it is observed that the piston movement in the sleeve can induce particle migration even before the material enters the cavity, an ALE (Arbitrary Lagrangian-Eulerian) formulation is also developed to include the piston movement in molding simulations. The ALE formulation is first compared with an Eulerian solution for the case of the piston driven flow problem. Then, the approach is applied to injection molding problems and the segregation inside the molded parts is studied.