Contents lists available at ScienceDirect International Journal of Heat and Fluid Flow journal homepage: www.elsevier.com/locate/ijhff CFD analysis of the erosion of a light gas stratification by means of a hot air jet in the MiniPanda facility A. Attavino ⁎ ,a , L. Koloszar a , M. Adorni b , P. Planquart a a Environment and Applied Fluid Dynamics Department, von Karman Institute for Fluid Dynamics, Waterloosesteenweg 72, Sint-Genesius-Rode B-1640, Belgium b BelV, Rue Walcourt 148, Brussels 1070, Belgium ARTICLEINFO Keywords: Light gases mixing Buoyancy driven flows Turbulence modeling ABSTRACT Following the Fukushima-Daiichi accident, many countries decided to strengthen the safety systems of their nuclear power plants in order to increase the capabilities of managing severe nuclear accidents. To achieve this goal detailed analysis of postulated design and beyond-design-basis accidents is essential. The main tools available for the analysis of these complex scenarios are Advanced Lumped Parameters (LP) and Computational Fluid Dynamics (CFD) codes. Currently, the most limiting factor in the application of these tools is their vali- dation. This activity is performed in support of the development and validation of a CFD model, built in the numerical environment of OpenFOAM, for hydrogen behavior in the containment of Light Water Reactors. For validationpurpose,numericalsimulationsoftheerosionofalightgasstratificationbymeansofahotairjethave been performed, and results of the simulations were analysed and compared against experimental results ob- tained by Ritterath (2012) in the MiniPanda facility. Two different scenarios have been considered, each characterized by a different value of the mass flow rate of the jet used to erode the stratification. 1. Introduction During a severe nuclear accident a huge amount of hydrogen can be generated by different sources in a Light Water Reactor (LWR) core. Hydrogen combustion might cause both mechanical and thermal loads. These loads could damage the equipment and, eventually, cause the failure of the reactor containment building leading to dispersion of radioactive material in the environment (OECD/NEA, 2014). There are many phenomena governing the behavior of containment in case of a severe accident, including gas (hydrogen, air and steam) mixing, steam condensation, turbulent transport and wall heat transfer. These phenomena are driven by the release of gases from the reactor. The release can happen in three different regimes: high momentum release (jet), buoyant low momentum injection (plume) and a transient regime in which the injection starts as a jet and turn into a pure plume after a certain distance (Yadigaroglu et al., 2003). Due to the simultaneous presence of all the phenomena mentioned above, the analysis of the thermal hydraulics behavior of the contain- ment during a severe nuclear accident is really complex. For this reason, advanced Lumped Parameters (LP) and Computational Fluid Dynamics (CFD) codes are fundamental tools for the analysis of LWR behavior during postulated design and beyond-design-basis accidents. In nuclear safety, the reliability of the analysis made with a computational tool is of prime importance. At present, the validation of codes is the major limiting factor in their application. The main issue for the validation of computational codes is the lack of adequate ex- perimentaldatawiththerequiredtemporalandspatialresolutions.This is especially true for the CFD grade experiments, where measurements should have a spatial and temporal resolution that can be compared with the CFD results. Therefore, in the past years many experimental project have been carried out (Gupta et al., 2015), in order to create a large database for the validation of CFD codes. The validation of a CFD code is a long procedure, and the effect of thedifferentphenomenathatareinvolvedinanuclearaccidenthaveto be analysed in an isolated way. The aim of this work is to develop a solver in the OpenFOAM simulation environment for simulations of the erosionofalightgasstratification.Thevalidationisdonebycomparing the simulation’s results against experimental data. The work includes the study of the effect of buoyancy in turbulence modeling, as well as, the influence of temporal discretization and mesh refinement. 2. MiniPanda experiments As a reference, the experiments performed in the MiniPanda facility in ETH Zürich (Ritterath, 2012) have been used, in fact, the spatial and temporal resolution of the measurements performed are suitable for the https://doi.org/10.1016/j.ijheatfluidflow.2019.04.012 Received 6 December 2018; Received in revised form 17 March 2019; Accepted 23 April 2019 ⁎ Corresponding author. E-mail address: andrea.attavino@vki.ac.be (A. Attavino). International Journal of Heat and Fluid Flow 78 (2019) 108409 0142-727X/ © 2019 Elsevier Inc. All rights reserved. T