Analyses of THAI 1 hydrogen deflagration using MELCOR code version 2.1 and 2.2

•Modelling and Simulation of two experiments using MELCOR code version 2.1 and version 2.2.•Benchmarking of the MELCOR code. Code-Experiment and Code-to-Code (different versions).•Assess of the MELCOR code capabilities to modelling and simulate hydrogen deflagration scenarios. The Fukushima Daiichi NPP (Nuclear Power Plant) accident pointed out the hydrogen explosion issue as one of the main problems that can affect the NPP containment integrity. During a severe accident scenario, the hydrogen combustion can occur and lead to containment integrity failure, since it generates local and global pressure and heat spikes. Such topic was analysed in several research programs addressed all around the world. An important series of test campaigns was done in OECD/NEA WGAMA (Organisation for Economic Co-operation and Development/Nuclear Energy Agency Working Group on Analysis and Management of Accidents) program called THAI (Thermal-hydraulics, Hydrogen, Aerosol and Iodine). The THAI goal is to simulate several phenomena related on hydrogen and Fission Product behaviour in the containment to obtain data relevant for the code benchmarking and validation. Therefore, theoretical analyses are needed, in order to obtain a reliable prediction of the accidental scenario. The facility allows to investigate safety relevant effects under thermal-hydraulics conditions of severe accidents. The experiments performed cover from hydrogen deflagration to iodine and aerosol behaviour under different thermal-hydraulics conditions. Three representative experiments were chosen from the THAI campaign to be modeled and simulated using the MELCOR code with versions 2.1 and 2.2 and compare the results with the experimental ones. This work aims to assess the MELCOR code capability pointing out on the limitation in simulating the hydrogen deflagration and underling possible method to reduce their effect on the simulate results. The benchmarks were addressed with old version of MELCOR however the new version presented slightly different results due to the modification in the parametric model and the default sensitivity coefficients.