Validation process against late phase conditions of the passive autocatalytic recombiner simulation code PARUPM as a standalone tool using experimental data from REKO-3 and THAI facilities

•PARUPM is a PAR simulation code that reproduce key phenomena related to the recombination process•PARUPM has undergone a thorough validation process as a standalone tool using data from REKO-3 and THAI experiments.•The code was enhanced to better simulate phenomena such as oxygen starvation and carbon monoxide recombination.•A CO poisoning criterion was introduced to the code to simulate the poisoning effect on the catalyst.•PARUPM is proposed for integration with thermal-hydraulic analysis codes for enhancing safety analysis in nuclear power plants. In case of a nuclear accident with core damage in a light water reactor, the oxidation of the fuel cladding and other materials could lead to the release of combustible gases (H2 and CO) to the containment building. To mitigate the potential risk of combustion of these gases, passive autocatalytic recombiners (PARs) have been installed in numerous nuclear reactors in Europe and worldwide. PARs recombine H2 and CO with O2 producing H2O and CO2, respectively, without an open flame. PARUPM is a code that simulates the behaviour of PARs using a physicochemical model approach. In the framework of the AMHYCO project (EU-funded Horizon 2020 project), which seeks to advance the understanding and simulation capabilities to support the combustion risk management in severe accidents, the code has been extensively enhanced and developed to simulate PAR operation with H2/CO/O2/steam mixtures. Alongside these new capabilities, the code needed a new validation process. In this paper, the process of validation of PARUPM as a standalone code is described. The validation for steady state conditions was achieved through comparison with REKO-3 experimental data while the transient conditions were compared with results obtained with the THAI test facility. A thorough analysis of the code capabilities was performed by comparing the numerical results with experimental data for a broad series of conditions, namely: a range of different input gas temperatures and concentrations, oxygen starvation, CO poisoning, etc.