The main objective of the present manuscript is to implement DEM simulations of powder spreading in an Additive Manufacturing process. A numerical sensitivity analysis is carried out in order to identify the contact law parameters that are the most impacting on powder bed quality. Tested parameters are surface energy, as well as sliding, rolling and restitution coefficients. It is seen that effective surface energy between powder particles is a key parameter to assess bed-forming ability, but that it is also necessary to consider rolling and sliding frictions for an accurate modeling of the spreading phenomena. In addition, our simulations allow a better understanding of the size segregation issues during the feeding and deposition stages. Finally, we describe the effect of the contact laws parameters on the heap profile. We identify various powder flow zones within the heap, and show that heap profile can be a predictor of powder bed quality.