The numerical results from the proposed methodology given in part 1 are presented. The sensitivity of the wall roughness of the pipe is analysed for both cases: subcooled liquid or two-phase mixture at the entrance of the pipe. A systematic comparison between models HEM and DEM and the experimental results shows that the metastability effects of the liquid phase must be taken into account even for long pipes. The predictions of the DEM are compared favourably with the experimental data of the mass flow rate as well as of the pressure profile obtained for simple-choked and double-choked steam-water flashing flows.