Natural and Forced Ventilation of Buoyant Gas Released in a Full-Scale Garage : Comparison of Model Predictions and Experimental Data

1
Natural and Forced Ventilation of Buoyant Gas
Released in a Full-Scale Garage Comparison of
Model Predictions and Experimental Data
Kuldeep Prasad, William Pitts, M. Fernandez, and
Jiann Yang
Fire Research Division National Institute of
Standards and Technology Gaithersburg, MD
20899. ICHS4 2011 e-mail kprasad_at_nist.gov
2
What is the problem?

• Fire safety in partially confined spaces
• Hydrogen Powered Systems
• Rising energy demands, Environmental degradation
• Potential for fires and explosions (unintended
  releases).
• Predicting temporally evolving concentration is
  challenging
• Unknown release rates, release location,
  orientation.
• Release duration (pressure of tank, size of
  release port)
• Unknown compartment leak sizes, locations.
• Effect of wind, thermal effects, forced
  ventilation.
• Predicting the effect of mitigation techniques
• Reduce envelope of flammable concentrations as
  quickly as possible.

3
Fire safety in partially confined spaces

• Motivation
• Literature Review Limited studies on buoyant gas
  released under an automobile in a realistic,
  full-scale garage (clutter).
• Methodology to predict hydrogen volume fraction
  in partially enclosed compartments Safety of
  hydrogen fueled applications ( improving codes
  and standards)
• Approach
• Detailed experimental, numerical and analytical
  modeling study to better understand hydrogen
  safety in partially enclosed compartments.
• Development of a simple and validated analytical
  model (based on compartment over-pressure).
• Study natural and forced dispersion of a buoyant
  gas.
• Compare / contrast results, Recommendations.

4
Experimental Setup
Garage 6.8 m x 5.4 m x 2.4 m Manual Door 2.8
m x 2.1 m Two Doors 1.5 m x 1.1 m Mid-size
passenger car parked centered over the release
location. Car windows rolled up and hood / trunk
closed.
Helium released under the car to simulate release
from a hydrogen fueled car. Flow rate monitored
with a dry test meter (5 kg of hydrogen over 4
hours). Release into the garage from a box 0.3 m
x 0.3 m x 0.15 m (diffuser , wiremesh, crushed
stone).
5
Experimental Setup Air-leakage Test
Helium volume fraction Thermal conductivity
sensors Estimate Air-leakage rates (ASTM
E779-03 standard) INFILTEC Model E3 Blower door
fan coupled with a digital micro-manometer DM4.
Pressure differential 10 Pa-70Pa. Mitigation
Strategy Used the built-in fan of an INFILTEC
duct Leakage tester to provide forced ventilation
in garage.
6
Flow stratification vs. well-mixed
Flow-field characterized by Fr Number.
Low-momentum jet buoyancy controlled
stratification. High momentum jet induces mixing
in the compartment. High pressure release under
an automobile turbulent mixing, break up into
multiple jets Hydrogen escapes from the wheel
wells and perimeter as multiple plumes (well
mixed compartment) Phenomena observed in CFD,
experiments.
7
Problem Formulation

• Model the flow through vents.
• Bernoulli equation.
• Pressure varies hydrostatically.
• Discharge coefficient.

Vertical pressure gradient inside compartment is
lower than the gradient outside the compartment.
Difference between these pressure gradients leads
to a buoyancy driven flow through the vents.
8
Conservation Equations
Conservation of Hydrogen in upper layer
Conservation of total mass in upper layer
Constraint Equation
Plume Modeling
Classical Plume Mixing Model self similar plume
solution
Buoyancy Flux
Effective origin
9
Prior work Reduced scale experiments
Quarter scale two-car residential garage

• 6.1 m x 6.1 m x 3.05 m
• Helium surrogate gas.
• Mass flow controller.
• Release of 5 kg of hydrogen.
• Time resolved concentrations.
• Idealized leaks, 3 ACH at 4 Pa
• Two square vents, 2.15 cm

Comparison with Analytical Model
10
Prior Work CFD modeling
Computational Domain

• NIST Fire Dynamics Simulator
• Low speed, chemically reacting fluid flow.
• Low Mach number, LES model
• 2nd order, multi-block rectilinear grid.

Comparison of CFD (symbols) with analytical model
(line)
11
Summary of tests conducted
Estimation of Leak Area (ASTM E779-03)
No. Test C n Q4Pa ACH ELA
1 Pressurization Test Side Door, Rear door un-sealed 0.022 0.64 0.0522 2.1 0.0202
2 De-pressurization Test Side Door, Rear door un-sealed 0.025 0.75 0.0699 2.8 0.0271
3 Pressurization Test Side Door, Rear door sealed 0.017 0.64 0.0423 1.7 0.0164
4 De-pressurization Test Side Door, Rear door sealed 0.024 0.72 0.0658 2.7 0.0255
5 Pressurization Test Rear Door, Side door un-sealed 0.031 0.50 0.0622 2.5 0.0241
6 De-pressurization Test Rear Door, Side door un-sealed 0.038 0.61 0.0894 3.6 0.0346
7 Pressurization Test Rear Door, Side door sealed 0.012 0.67 0.0305 1.2 0.0118
8 De-pressurization Test Rear Door, Side door sealed 0.013 0.86 0.0438 1.8 0.0170
Natural and Forced Ventilation Tests with
Automobile
No. Release Rate (m3/s) Release Duration (s) Natural / Forced Forced Flow Rate (m3/s) Fan Start Time (s)
1 0.004468 14396 Natural -
2 0.004326 13622 Natural -
3 0.004434 3628 Forced 0.0910 125
4 0.004239 3599 Forced 0.0922 101
5 0.004273 3618 Forced 0.1066 101
6 0.004283 3597 Forced 0.1071 0
12
Natural Ventilation Tests
Garage Measurements
61.0 cm 30.5 cm
Car Measurements
Under car center
Top of engine
Wheelwell
Engine Compartment
DomeLight
Trunk
13
Forced Ventilation Tests
Garage Measurements
Car Measurements
Under car center
Top of engine
Wheelwell
Engine Compartment
14
Summary and Conclusions

• A detailed experimental, numerical and analytical
  modeling study to better understand and improve
  the safety of hydrogen fueled applications in
  passively and actively ventilated spaces.
• Validation of analytical model with CFD and
  reduced scale expt. (allows it to be used for
  improving hydrogen safety codes and standards)
• Models results over-predicted the experimental
  data by 0.4 for natural ventilation conditions
  and 1.0 for forced ventilation conditions.
• Parametric studies to understand the effect of
  release rates, vent size and location on the
  predicted helium volume fraction.
• Analytical model does not predict the pockets of
  buoyant gas at concentrations that are
  significantly higher than the LFL, does not
  predict the seepage of helium inside the vehicle.

15
Effect of Input Parameters
Effect of Vent Height
Effect of Compartment Volume
16
Effect of Input Parameters
Effect of Upper Vent Area
Effect of Lower Vent Area
17
Prior Work Wind Driven Ventilation
Assisting Wind Flow
Weak / Strong Opposing Wind Flow
18
High-pressure Release Exit Conditions

• Ideal-gas behavior - pressure less than 17.4
  MPa.
• Gas compressibility effect at higher pressure.
• Abel-Noble equation of state
• High pressure release from a storage tank.
• Flow rate reduces with stagnation pressure.
• Choked flow at jet exit.
• Exit pressure is greater than atmospheric
  pressure.
• Flow expansion through series of expansion
  shocks.
• Choked sonic release lasts until
• Sub-sonic flow conditions beyond this point
• Exit pressure equal to atmospheric pressure.
• Use of Isentropic Flow Relationships to model
  tank blow down time and volumetric flow rate.
  (Shapiro)

19
Forced Venting of Hydrogen Results
20
Effect of Hydrogen Release Rate
Hydrogen Volume Fraction
Height of the Interface
Volumetric Flow Rates
Compartment Overpressure
21
Wind Driven Ventilation-Steady State Results
22
Time required to empty a compartment
Wind assisted venting
Buoyancy driven flow
23
Release as a Distributed Source

• Fully mixed hydrogen air mixture in
  compartment.
• Goal predict hydrogen concentrations in
  compartment
• Pressure varies hydrostatically with depth. Vent
  flows driven by pressure difference.

• Release of hydrogen under an obstruction.
• Cluttered environment, Multiple plumes.
• Buoyant gas mixes rapidly with surrounding air..

24
Problem Formulation
Conservation of hydrogen mass
Constraint Equation
25
Effect of Vent Area, Location
Multiple Vents
26
Conclusions and Summary

• Natural and wind driven ventilation of hydrogen
  released in an accidental manner in a partially
  enclosed compartment.
• Development of simple analytical models
• Validated with reduced scale experiments.
• Validated with full scale detailed CFD
  simulations.
• Effect of hydrogen release rate.
• Effect of vent cross-sectional area, distance
  between vents, multiple vents, location of vents.
• Role of assisting and opposing wind flows
• Forced ventilation, buoyancy driven flow, thermal
  effects.
• Effect of surrogate gas (helium).
• Time to empty a compartment filled with hydrogen
  gas.

27
(No Transcript)
28
High Pressure Release and Dispersion of Hydrogen
in a Partially Enclosed Compartment
Kuldeep Prasad, Thomas Cleary and Jiann Yang
Fire Research Division Engineering Laboratory
National Institute of Standards and
Technology Gaithersburg, MD 20899. Fuel Cell and
Hydrogen Energy 2011 Corresponding Author
kprasad_at_nist.gov
29
Current and Future Technologies

• Hydrogen powered systems
• Hydrogen Energy carrier for future vehicles.
• Driven by rising energy demands.
• Environmental degradation problem.

30
What is the problem?

• Current technologies require
• High pressure storage of hydrogen (70 MPa).
• Acceptable levels of vehicle driving range,
  storage volume and weight requirements.
• Risk associated with high pressure releases
• Damage to storage tank, piping or PRD failure.
• Dispersion in partially enclosed compartments.
• Effective mitigation techniques and requirements.
• Support standard and code development.

31
Prior Work Reduced scale experiments
Quarter scale two-car residential garage

• 6.1 m x 6.1 m x 3.05 m
• Helium surrogate gas.
• Mass flow controller.
• Release 5 kg of hydrogen.
• Time resolved concentrations.
• Idealized leaks, 3 ACH at 4 Pa
• Two square vents, 2.15 cm

Comparison with Analytical Model
32
Justification of well-mixed assumption

• Flow-field in compartment characterized by
  Froude Number.
• Low-momentum jet buoyancy controlled
  stratification.
• High momentum jet induces mixing in the
  compartment.
• High pressure release under an automobile
  turbulent mixing, break up into multiple jets
  Hydrogen escapes from the wheel wells and
  perimeter as multiple plumes (distributed
  sources).
• Phenomena observed in experiments performed in
  full scale garages as well as CFD simulations.

33
Problem Formulation

• Model the flow through vents - Bernoulli
  equation.
• Pressure varies hydrostatically with depth.

Vertical pressure gradient inside compartment is
lower than the gradient outside the compartment.
Difference between these pressure gradients leads
to a buoyancy driven flow through the vents.
34
Conservation Equations
Conservation of Hydrogen
Constraint Equation
Design of Idealized Vents
Leak rates described in terms of air changes
per hour (ACH). Exchange across enclosure
boundary - Effective Leak Area (ELA) of
enclosure related to ACH varies widely for
garages (recommended value ACH3)
35
40 Mpa, 5 kg tank, 1 mm release port Jet Exit
Conditions
Tank Pressure
Exit Velocity
36
Release and Dispersion
Compartment Overpressure
Volume Fraction
37
Height of Interface
Vent Flow Rates
38
40 Mpa, 5 kg tank, 6 mm release port Jet Exit
Conditions
Exit Velocity
Tank Pressure
39
Release and Dispersion
Compartment Overpressure
Volume Fraction
40
Height of Interface
Vent Flow Rates
41
Compartment overpressure vs. Diameter of release
port
42
40 Mpa, 5 kg tank, 1 mm release port Forced Flow
Rate 0.1 m3/s
Compartment Overpressure
Volume Fraction
43
Height of Interface
Vent Flow Rates
44
Forced Flow Rate vs Peak Volume Fraction
Forced Flow Rate vs Duration of flammable mixture
45
Conclusions and Summary

• Developed a simple analytical model to predict
  the risk associated with accidental release of
  hydrogen from a high-pressure system in a
  partially ventilated compartment.
• Assumed that the hydrogen released under an
  automobile mixed rapidly with the surrounding
  air.
• Analytical model for natural and forced mixing
  and dispersion of hydrogen released in a
  compartment.
• Ventilation of the compartment occurs through two
  idealized holes in the compartment walls (ACH
  varied between 1-5).
• Examine conditions that lead to major damage of
  the compartment due to overpressure.
• 6 mm diameter release port Significant damage
• 1 mm diameter release port Cosmetic damage
• Forced ventilation is a viable technique for
  reducing dangerous levels of hydrogen
  concentration in compartment.

46
Conclusions and Summary (cont.)

• Model can be used to provide design guidelines
  for forced ventilation requirements in a
  compartment
• Simple analytical models have been
• Validated with reduced scale and full scale
  experiments.
• Compared with detailed CFD simulations
• Effect of hydrogen release rate, vent
  cross-sectional area, distance between vents,
  multiple vents, location of vents.
• Role of assisting and opposing wind flows.
• Forced ventilation, buoyancy driven flow, thermal
  effects.
• Effect of surrogate gas (helium).
• Time to empty a compartment filled with hydrogen
  gas.

47
Contact Information
Kuldeep Prasad, Thomas Cleary and Jiann Yang
Corresponding Author Email kuldeep.prasad_at_nist
.gov Fire Research Division Engineering
Laboratory National Institute of Standards and
Technology Gaithersburg, MD 20899.