I noticed that when running an inflow-outflow simulation using the generated 3D mesh (base.rea), the simulation blows up. This doesn't happen if the 3D mesh is obtained extruding base2d.rea with n2to3. From a first investigation it might be due to what is written in the line O of the boundary conditions in the .rea and to an update in the call of the routine BCNEUTR.