NUMERICAL STUDY OF SPONTANEOUS IGNITION OF PRESSURIZED HYDROGEN RELEASE INTO AIR

1
NUMERICAL STUDY OF SPONTANEOUS IGNITION
OFPRESSURIZED HYDROGEN RELEASE INTO AIR

• B. P. Xu1, L. EL Hima1, J. X. Wen1, S. Dembele1
  and V.H.Y. Tam2
• 1Faculty of Engineering, Kingston University
• Friars Avenue, London, SW15 3DW, UK
• 2EPTG, bp Exploration, Chertsey Road,
  Sunbury-on-Thames, TW16 7LN, UK

2
Analysis of Hydrogen Accidental Database Compiled
by Kingston University
3
Postulated Ignition Mechanisms(Astbury
Hawksworth, 2005)

• Reverse Joule-Thomson effect
• Static electricity
• Sudden adiabatic compression
• Hot surface ignition
• Mechanical friction and impact
• Diffusion ignition

Self-ignition and explosion during discharge of
high-pressure hydrogen, Toshio Mogi, Dongjoon
Kim, Hiroumi Shiina, Sadashige Horiguchi, J of
Loss Prevention, 2007.
4
Diffusion Ignition(Wolanski and Wojcicki, 1973)
Temperature

• Sudden rupture of a pressure boundary
• Shock wave (primary)
• Shock-heated air or oxidizer (behind the shock
  wave)
• Cooling hydrogen flow acceleration or
  divergence
• Formation of combustible mixture at contact
  surface molecular diffusion
• Ignition (after a delay time)

Schlieren Density
5
Mixing at Contact Surface

• Mass and energy exchange between shock-heated air
  and cooled hydrogen through molecular diffusion
• Molecular diffusivity is inversely proportional
  to local pressure and proportional to local
  temperature
• The thickness of the contact surface extremely
  thin and increasing with release time

For the release case of 100 bar through a 1mm
hole at t3us
6
Ignition its Delay Time

• Mixing time
• the time for the temperature of the mixture
  to reach the autoignition temperature

• Chemical delay time
• due to the slow hydrogen combustion rate
  under low temperature

An increase in the release pressure will lead to
a rapid decrease in both mixing time and chemical
delay time
7
Numerical Methods

• 2-D Unsteady Compressible Navier-Stokes equations
  solved with an ILES method
• Detailed chemical-kinetic scheme - 8 reactive
  species and 21 elementary steps third body and
  fall off behavior considered (Williams 2006)
• Multi-component diffusion approach for mixing -
  thermal diffusion
• ALE numerical scheme convective term solved
  separately from diffusion terms
• In Lagrangian stage, 2nd-order Crank-Nicolson
  scheme 2nd-order central differencing
• In rezone phase, 3rd-order Runge-Kutta method
  5th - order upwind WENO scheme

Numerical diffusion resulting from the use of
lower order schemes could lead to overprediction
of the contact surface thickness, and hence
artificially increase the chance of autoignition.
8
Problem Description

• Three hole sizes
• 1mm, 3mm, 5mm
• Three release Pressures
• 100 bar, 200 bar, 300 bar
• Non-slip and adiabatic wall boundary
• Minimum cell size
• 10 microns
• Only early stage of release simulated

9
Overview of the Under-expanded jet
1. Mach Number
2. Temperature
3. Mass Fraction
3. Density Schlieren
10
The Very Early Release Moment
1. Axial Velocity
2. Temperature
3. Density Schlieren
11
Changing Patterns for the release case of 300 bar
through a 3mm hole at t9us
12
Contour of OH for the release case of 300 bar
through a 5mm hole at t35us
13
Contour of Mach Number for the release case of
300 bar through a 5mm hole at t35us
14
Contour of Temperature for the release case of
300 bar through a 5mm hole at t35us
15
Contour of Mass Fraction for the release case of
300 bar through a 5mm hole at t35us
16
Density Schlieren for the release case of 300 bar
through a 5mm hole at t35us
17
Release through a 1mm, 3mm and 5mm holes
18
Grid Sensitivity - fine mesh (10 microns), coarse
mesh (20 microns)
19
Concluding Statement
The release of highly pressurised hydrogen into
air could lead to spontaneous ignition depending
on pressure, hole sizes, etc.   Current work
demonstrated the conditions leading to
autoignition. A flame was found to be sustained
over a period of 50 ?s and still stable when the
simulation was terminated due to limitation of
computer power.  Further work is underway to
establish whether this flame could be maintained,
leading to a jet fire, fire ball or an explosion.
20
Conclusions

• Diffusion ignition has been numerically proved to
  be a possible mechanism for spontaneous ignition
  of a sudden high pressure hydrogen release direct
  into air
• Ignition first occurs at the tip of a very thin
  contact surface and then moves downward along the
  contact surface
• The pressure at the contact surface is higher
  than the ambient pressure this high pressure
  produces a high heat release rate to overcomes
  the flow divergence and sustain the very thin
  flame
• As the jet develops downstream, the front of the
  contact surface becomes distorted
• The distorted contact surface can increase the
  chemical reaction interface and enhance the
  mixing process, and it may be the key factor for
  the final diffusive turbulent flame
• Finally, an increase in the release pressure or
  the hole size can facilitate the ignition.

21
ACKNOWLEDGEMENT

• B P XUs Post-doctoral Fellowship is funded by
  the Faculty of Engineering at Kingston
  University.
• The authors would like to acknowledge EU FP6
  Marie Curie programme for funding hydrogen
  research at Kingston University through the
  HYFIRE (HYdrogen combustion in the context of
  FIRE and explosion safety) project.
• BP and HSL are supporting groups for HYFIRE.