IA-HySafe Standard benchmark exercise
SBEP-V21: Hydrogen release and
accumulation within a non-ventilated ambient
pressure garage at low release rates
A.G.Venetsanos1, I. Tolias1, D. Baraldi7, S. Benz5, B.
Cariteau2, J. Garcia3, O.R. Hansen4, C. Jäkel6, S. Ledin8, P.
Middha4, E.A. Papanikolaou7
1 Environmental Research Laboratory, National Centre for Scientific Research Demokritos
(NCSRD), 15310 Aghia Paraskevi, Attikis, Greece, [email protected]
2 C.E.A. Saclay, D.E.N., D.M.2S., S.F.M.E., Laboratoire d’Etude Expérimentale des Fluides,
91191 Gif/Yvette cedex FRANCE
3 Escuela Técnica Superior de Ingenieros Industriales, Universidad Politécnica de Madrid
(UPM), José Gutiérrez Abascal, 2, E-28006 Madrid, Spain
4 GEXCON AS, Fantoftvegen 38 Box 6015 Postterminalen N-5892 BERGEN Norway
5 IKET, KIT, Postfach 3640, 76021 Karlsruhe, Germany
6 Forschungszentrum Juelich (FZJ), 52425 Juelich, Germany
7 Joint Research Centre of the European Commission (JRC), Institute for Energy, 1755 ZG
Petten, The Netherlands
8 Health and Safety Laboratory (HSL), Harpur Hill, Buxton, Derbyshire, SK17 9JN, UK
ICHS-4, 12-15 Sept. 2011, San Francisco, USA
CONTENTS





Introduction & Scope
Description of CEA Garage Test 5
Modelling strategy
Results & discussion
Conclusions
ICHS-4, 12-15 Sept. 2011, San Francisco, USA
2
Intro & Scope

The Research Committee activities of IAHySafe include:


SBEPs to


the continuation of SBEPs first introduced within
HySafe EC-NoE
validate, inter-compare and further develop
existing Computational Fluid Dynamics (CFD)
codes and models in predicting hydrogen related
release, dispersion and combustion phenomena
Focus in the present work:

Understanding and prediction of the dispersion /
accumulation of hydrogen releases in confined
spaces under low release conditions
ICHS-4, 12-15 Sept. 2011, San Francisco, USA
3
CEA Garage Test-5
Re = 115
Garage x-dimension (mm)
5760
Garage y-dimension (mm)
2960
Garage z-dimension (mm)
2420
x release (mm)
y release (mm)
z release (mm)
Exit diameter (mm)
Volumetric flow rate - STP
(NL/min)
He mass flow rate (g/s)
-2880
1480
220
29,7
Garage Temperature T (°C)
24,1
Exit velocity (m/s)
Release Direction
Release Type
Release duration (s)
0,47
Upwards
Continuous
3740
Released volume - STP (NL)
1122
18
0,054
He released mass (gr)
200,28
Target concentration (%)
2,94%
Total measurement time (s)
90440
ICHS-4, 12-15 Sept. 2011, San Francisco, USA
4
Modelling strategy
Participa
nt/
Code
Turbule
nce
model
Number of
computational
cells
Convective
terms
Transient
terms
Vent
model
Molecular
diffusivity
(m2/s×10-5)
FZJ/CFX
SAS-SST
212272 (half
garage)
2nd order
2nd order
backwards
Euler
Opening
8.0
k-ε
Release: 30888
Diffusion: 4560
2nd order
Opening
2.0
k-ε
Release: 44352
Diffusion: 7200
2nd order
Extension
2.0
2nd order
backwards
Euler
Opening
1.86
GEXCON/
FLACS
GEXCON_
revised/
FLACS
HSL/CFX
SST
88840
High
resolution
scheme (2nd
order)
JRC/CFX
Laminar
Release: 101401
Diffusion: 70429
Outside: 5173
2nd order
2nd order
backwards
Euler
2m×3m×
3m
extension
11.7
k-ε
58905
Upwind 1rst
order
ALE scheme
Opening
7.5
k-ε
53760
Outside: 3360
SMART 3rd
order
1rst order
fully implicit
1 m × 2.96
m × 2.42 m
extension
5.65
Laminar
125516
2nd order
1rst order
fully implicit
Opening
5.65
Laminar
-
1 m × 2.96
m × 2.42 m
extension
5.65
KIT/
GASFLOW
NCSRD/
ADREAHF
UPM/
FLUENT
UPM_revi
sed/
FLUENT
ICHS-4, 12-15 Sept. 2011, San Francisco, USA
5
Results – Mass balance
ICHS-4, 12-15 Sept. 2011, San Francisco, USA
6
Results – CPU times
Participant/
Code
Compressibility
Simulated period (s)
CPU time
Threads
Computer type
FZJ/CFX
Fully
compressible
20000 s
15d, 21h
3
Intel i7 860 CPU
Low Mach
100000 s
1d
1
Linux-PC, Intel Xeon
W3550, 3.07 GHz
Low Mach
20000 s
5.5h
4
Quad core
HSL/CFX
Fully
compressible
15000 s (coarse
mesh) 12000 s
(fine)
18d, 5h
(coarse) 19d,
15h (fine)
2
Windows XP SP3 64bit, 2 Intel Xeon
JRC/CFX
Fully
compressible
15000 s
9d, 20h
8
3.25 GHz
KIT/
GASFLOW
Low Mach
20000
2d, 15h
1
Suse Linux 11.2, i7950, 3.06GHz
NCSRD/
ADREA-HF
Fully
compressible
7080
10d
2
Windows 7, i7 M620
CPU at 2.67 GHz
UPM/
FLUENT
Fully
compressible
20000
13d
4
Windows 7, 2 Intel
Xeon Quad Core CPUs
at 2.4 GHz
GEXCON/
FLACS
GEXCON_rev
ised/
FLACS
ICHS-4, 12-15 Sept. 2011, San Francisco, USA
7
Results - Concentrations
Sensor name
P1N3
P1N2
P1N1
M1N5
M1N4
M1N3
M1N2
M1N1
z (mm)
2370
2135
1900
1575
1260
945
630
315
ICHS-4, 12-15 Sept. 2011, San Francisco, USA
8
Results - Concentrations
P1N3, z = 2.37 m
ICHS-4, 12-15 Sept. 2011, San Francisco, USA
9
Results - Concentrations
P1N2, z = 2.135 m
ICHS-4, 12-15 Sept. 2011, San Francisco, USA
10
Results - Concentrations
P1N1, z = 1.9 m
ICHS-4, 12-15 Sept. 2011, San Francisco, USA
11
Results - Concentrations
M1N5, z = 1.575 m
ICHS-4, 12-15 Sept. 2011, San Francisco, USA
12
Results - Concentrations
M1N4, z = 1.26 m
ICHS-4, 12-15 Sept. 2011, San Francisco, USA
13
Results - Concentrations
M1N3, z = 0.945 m
ICHS-4, 12-15 Sept. 2011, San Francisco, USA
14
Results - Concentrations
M1N2, z = 0.63 m
ICHS-4, 12-15 Sept. 2011, San Francisco, USA
15
Results - Concentrations
M1N1, z = 0.315 m
ICHS-4, 12-15 Sept. 2011, San Francisco, USA
16
CONCLUSIONS



Relatively good predictions of He
concentrations
Significant reduction in CPU time with low
Mach number solvers
Significant effect of extending the
computational domain far enough beyond
the confined space opening.
ICHS-4, 12-15 Sept. 2011, San Francisco, USA
THANK YOU FOR
YOUR ATTENTION
ANY QUESTIONS
ICHS-4, 12-15 Sept. 2011, San Francisco, USA