Validation and application of the REKO-DIREKT code for the simulation of passive auto-catalytic recombiners (PARs) operational behaviour

In order to reduce the accumulation of hydrogen and thus to mitigate the risk of a combustion, passive auto-catalytic recombiners (PARs) have been installed within LWR containments in many countries. The severe hydrogen combustion events in the recent Fukushima–Daiichi accident are likely to imply an increased demand in upgrading nuclear power plants with PARs. Numerical simulation is an important tool in order to assess PAR operation during a severe accident in terms of efficiency and proper installation. For the quite challenging boundary conditions during a severe accident, including e.g. low oxygen amount, high steam amount, presence of carbon monoxide, advanced numerical PAR models are required. The REKO-DIREKT code has been developed in order to provide a PAR model capable to simulate complex PAR phenomena and at the same time being suitable for implementation in thermal hydraulics codes. The development of REKO-DIREKT was supported by small-scale experiments performed at JÜLICH in the REKO facilities. These facilities allow to study PAR related single phenomena such as reaction kinetics under different conditions including variation of steam, oxygen and carbon monoxide (REKO-3) and the chimney effect (REKO-4). Recently, the code has been validated against full-scale experiments performed in the THAI facility at Eschborn/Germany in the framework of the OECD/NEA-THAI project. By this, the code has proven its applicability for different PAR designs and for a broad range of boundary conditions (pressure of up to 3 bar, steam amount up to 60 vol.%, low-oxygen conditions). REKODIREKT has been successfully implemented in the commercial CFD code ANSYS-CFX as well as in the LP code COCOSYS (GRS, Germany). The paper gives an overview of the basic code features and development steps. Different validation steps are presented from stand-alone application to the analysis of a full experimental transient by means of code coupling with the CFD code ANSYS CFX 15. The consistent representation of all test parameters underlines the good predictive capabilities of the modelling approach.