The Accident Source Term Evaluation Code (ASTEC) and Methods of Estimation of Leakages and Consequences of Releases (MELCOR) system codes are used in Tractebel to perform source term assessments and plant Thermal Hydraulic (TH) analyses in case of Severe Accident (SA) for the Belgian Nuclear Power Plants (NPPs).These system codes were developed through intense and extensive validation activities which were mainly based on the verification of the capability of implemented physical models to reproduce the experimental results of Separate Effect Tests (SETs) and Coupled Effect Tests (CETs).However, the SETs, and in less extension the CETs, do not take into account interactions and possible synergistic or antagonistic effects which can arise among the different physical and chemical processes occurring during a SA. Thus, if a code is able to reproduce correctly most of SETs or CETs, this does not mean that it can ensure the same quality of results for the more realistic and complex Integral Effect Tests (IETs).Nevertheless, some contraindications exist also for the IETs, because of the scaling factor effect which, being difficult to quantify, could lead to development of models replicating correctly the experimental data but that are not totally reliable for reactor scale applications.In addition, waiting for the dismantling of Fukushima Dai-ichi units 1–2 and 3, which can provide new information in the field of the core degradation phenomena, in the present only data on in-vessel core melt progression for the Three Mile Island, Unit 2 (TMI-2) accident are available.Therefore, apart from simulating the TMI-2 accident scenario, there is no other way to assess models of in-vessel phenomena (H2 production, corium pool/ debris formation and relocation, etc) implemented in the codes on reactor scale.As a consequence, there are still uncertainties in the capability of these integral tools to predict the progression of postulated severe accident transients in a NPP, notably core degradation and ex-vessel phenomena.In order to investigate the differences among the physical models adopted in MELCOR 2.2 and ASTEC V2.2 codes, and evaluate the impact that they could have on the assessment of Severe Accident Management Strategies, a crosswalk study was carried-out.This paper describes the first phase of this work, which consists of a detailed comparison between these two SA tools on a well-defined plant (PWR1000-Like) with prescribed boundary and initial conditions.Furthermore, to avoid additional and unwanted sources of discrepancies between code calculations, the two reactor models were developed in parallel following the recommendations suggested by developers for core and Reactor Coolant System (RCS) nodalization.The comparative study is focused on the TH of the RCS and in-vessel phenomena, for two different accident scenarios (SBLOCA and CSBO) up to the moment of lower-head failure.Ex-vessel behaviour will be examined in the second phase of this study.The crosswalk results obtained have shown some differences and similarities on reproducing TH in the RCS and In-vessel phenomena.Especially for these latter, the discrepancies predominate mainly due to how the codes treat corium behaviour, while thanks to the harmonisation of the initial steady-state and boundary conditions, the discrepancies on the prediction of the RCS behaviour have been minimized, at least before significant core degradation takes place.