After the Fukushima accident, “st ress- test” activities carried out worldwide pointed out the need to study additional accident management measures to deal with prolonged Station Black Out (SBO) scenarios. Without any operator actions, a total loss of the secondary side heat sink leads to core uncovery, to core damage and ultimately to a melt -down scenario. The international NEA/OECD PKL -3 project has addressed the efficiency of possible accident management actions to re- establish core cooling by experiments at the PKL and PMK test faciliti es. The experiments at the PKL test facility were focused on the efficiency of the counter measures to re-establish the core cooling and the performance of core exit temperature measurements during the core heat up phase. In this type of scenarios, the cor e power has been significantly reduced; there are no break locations and only either sporadic or temporary openings of relief or safety valves. In this situation, sources of momentum are much smaller than in typical LOCA or SBLOCA events. Since best estimate system codes were mainly developed to simulate LOCA scenarios (in their full spectrum) the performance of system codes and the general guidelines followed to simulate PWR power plants are called into question. In this paper, RELAP5 simulations of two SB O experiments of the PKL facility are presented. Firstly, a post -test analysis of Test H2 -1 is presented and specific guidelines on modelling are given. Secondly, these guidelines are applied to the PKL nodalization to blindly simulate Test H2 -2. Finally, a combined post -test for both experiments has been performed. The RELAP5 system code has been assessed for its use in this type of scenarios.