In a 3D contact problem, after defining the boundary conditions (free surface and contact conditions) and selecting Lagrangian Remeshing method, the only interpolation options are “Linear velocities, constant pressure”. Is this true for the contact problem? When can I use other schemes such as mini-element or quadratic velocities? Is this limitation to have a realistic solution time for 3d contact problems? When one has a linear interpolation for the velocity, we have the same number of unknowns for the velocity and coordinates. In other words, there is a velocity vector for each node and the node moves with that velocity. If another interpolation would be selected for the velocity, there would be more velocity vectors than available nodes. For several simulation cases, this does not make any trouble, e.g. for steady extrusion flows. However, for contact simulations, this could have some consequences: a typical side effect indeed would be that one may have a non-zero velocity in the middle of a face while there would be no corresponding node motion. In other words, it means an entry or exit of material. Furthermore, in contact simulations, contact is detected only at actual mesh nodes where a velocity can be indeed cancelled via the mechanism of penalty. If one has a concave border, one could easily imagine a situation where all nodes are detected in contact but the central velocity would perhaps never see the contact. In summary, having an equal number of material nodes and velocity vectors is the best option.