Cerfacs Enter the world of high performance ...

Numerical study of hydrogen deflagrations in confined spaces: interactions with walls and flame acceleration mechanisms

MEGEP (Mécanique, Energétique, Génie civil & Procédés) – [Subject to defense authorization]

The global energy demand is continuously increasing, primarily met through the combustion of fossil fuels. This process releases greenhouse gases that contribute to global warming. Consequently, the current trend of climate change inquires for a rapid and ambitious energy transition. In this context, hydrogen is gaining momentum as an alternative energy carrier capable of replacing fossil fuels. It has the potential to 1) create a virtuous cycle for future carbon-neutral electricity grids by acting as a buffer for the intermittency of non-dispatchable renewable generation, and 2) decarbonize various sectors that rely on power generation, such as aviation. Nonetheless, the deployment of hydrogen in novel applications raises a significant set of safety concerns due to its specific properties, including wide flammability limits, low minimum ignition energy, high diffusivity, enhanced burning velocity, and strong propensity to detonate. Unintended releases of hydrogen through loss-of-containment scenarios therefore warrant thorough investigation. Computational Fluid Dynamics (CFD) places as a powerful tool for the risk assessment of hydrogen-related hazards, offering a safer alternative to experimental testing and providing detailed spatial and temporal data for any physical quantity within the computational domain. This thesis undertakes a numerical investigation of hydrogen explosion scenarios using CFD and structures around two main research axes: the treatment of wall boundaries in hydrogen-air deflagrations —a subsonic mode of combustion, propagating at several meters per second— with a focus on both chemical reactivity and heat transfer at the walls; and the investigation of Flame Acceleration (FA) mechanisms that can potentially trigger detonation —a supersonic mode of combustion that is much more destructive— through the deflagration-to-detonation transition phenomenon. Particular attention is paid to validating the Large Eddy Simulation (LES) framework for such scenarios. In the first part of this work, it is demonstrated that the commonly used inert wall assumption in flame-wall interaction studies involving detailed chemistry is inappropriate, as it leads to non-physical heat generation at walls where flame quenching is expected. Simplified wall-chemistry models should therefore be employed, such as the one proposed herein. Furthermore, the influence of heat losses on FA is examined. Results indicate that the dominant effect of heat loss on FA originates from the burnt gas side, while local flame quenching plays a secondary role. Given the short characteristic timescales of deflagration scenarios (on the order of milliseconds), cold isothermal wall boundary conditions represent the most physically realistic thermal treatment for safety-related simulations. The second part of the thesis focuses on application cases, examining deflagrations in confined geometries featuring either transverse jets or staggered solid obstacles. The LES framework is shown to accurately reproduce the flame dynamics observed in corresponding laboratory-scale experiments, thereby paving the way for the future deployment of LES-based risk assessment methodologies in industrial contexts.

Jury

Mme Valeria Di Sarli	Università di Bologna	Reviewer
M. Andrea Gruber	NTNU Trondheim/SINTEF	Reviewer
Mme Carmen Jiménez	CIEMAT Madrid	Examiner
M. Peter Lindstedt	Imperial College of London	Examiner
M. Thierry Poinsot	IMFT/CERFACS	Thesis supervisor
M. Quentin Douasbin	CERFACS	Thesis co-supervisor
M. Lucien Gallen	AIRBUS Protect	Invited member