Industrial powder extrusion and calendering processes have already been extensively studied in the past decades. However, the growing interest in dry granulated extrudates sets new challenges due to the decreased presence of liquid solvent usually added to lubricate these processes. Although some approaches already exist to understand the phenomena inherent to this type of flow, few of them are based on a discrete approach, while many of them favor a classical continuous representation of the problem [1]. The latter does not provide a perspective on the macroscopic properties of the flow, making it impossible to reproduce a flow blockage within the extrusion die resulting from the purely geometric properties of the powder particles forming the flow. We propose here a DEM model to address such flows in order to encapsulate the optimum operating range of a ram extruder for dense granular flow with low solvent content. The particles are tracked by a Lagrangian approach, their interactions being obtained by solving Newton's second law for each particle. The contact resolution is performed with a smooth contact dynamic model, allowing particle interpenetrations. A discrete cohesive interaction force is used to represent the polymeric bridges between the particles induced by the solvent [2]. The objective is to accurately represent the flow in the die, in particular to distinctively identify the obstruction factors (cold sintering) and the lubricative impact of the solvent. The measurement of the residual mechanical stresses and the possible damage of the contact network in the treated material allow the prediction of its mechanical integrity [3].