Skip to content

Numerical Modeling of a Moderate Hydrogen Leakage in a Typical Two-vented Fuel Cell Configuration


Numerical results are presented from two direct numerical simulations (DNS) where a moderate hydrogen leakage is modeled in a typical two-vented fuel cell configuration. The study mimics one of the experimental investigations carried out on the 1 m3 enclosure with a leak flow rate of 10.4 Nl.min−1 [1]. The injection dimensionless Richardson number is at the order of unity and thus characterizes a plume flow which becomes turbulent due to gravitational accelerations. Two large exterior regions are added to the computational domain to model correctly the exchange between the in/out flows at both vents and the outer environment. Two meshes are used in this study; a first consisting of 250 million cells, while the second has 2 billion cells to ensure the fine DNS resolution at the level of Kolmogorov and Batchelor length scales. The high performance computation (HPC) platform TRUST is employed where the computational domain is distributed up to 5.104 central processing unit (CPU) cores. A detailed description of the flow structure and the hydrogen dispersion is provided where the sharp effect of the cross-flow on the plume is analyzed. Comparisons versus the experimental measurements show a very good agreement where both the bi-layer Linden regime and the maximal concentration in the top homogeneous layer are correctly reproduced by the DNS. This result is extremely important and breaks the limitations shown previously with statistical RANS approaches and LES models. This study can be considered as a good candidate for any further improvements of the theoretical industrial plume models in general, and for the estimation of the non-constant entrainment coefficient in particular.

Related subjects: Safety
Countries: France

Article metrics loading...

This is a required field
Please enter a valid email address
Approval was a Success
Invalid data
An Error Occurred
Approval was partially successful, following selected items could not be processed due to error