First Flames:

Burning, Turbulence and Buoyancy

Jonathan Dursi (CITA), Frank Timmes (LANL),
Mike Zingale (SUNY SB)
 Flame Ignition
• Highly nonlinear,
transient
phenomenon

• Interaction of
nuclear reactions,
conductivity,
hydrodynamics

• Difficult to model

Nigel Gorbold, Univ. Colorado

 Ignition

• Turbulent
Niemeyer, Hillebrandt &
Woosley (1996):

conditions success of explosion

can depend sensitively
on number, location of

• Many potential ignition points

`flashpoints’
• Important to
understand where/
when points do
ignite
Early Burning
• Ignition can occur on ~cm scales
• Not resolved by large-scale
simulation
• For initial conditions in large-
scale simulations, need to map
• turbulence →ignition points
• ignition → early flame bubbles
 Outline
• Some notes on ignition
• One-zone
• Flame (ongoing)
• Detonation
• Early burning: what happens next?
• Characteristic size of flame bubbles
• Flame model?
 Flame vs Detonation
• Flame/Deflagration

• Subsonic

• Heat propagates by
conduction

• Detonation

• Supersonic

• Shock-driven heating
 Our contributions

• Dursi & Timmes 2006:

• One-zone ignition times
• Very hard to ignite detn at low densities:

need much more than local supersonic

flow. Geometric constraint.
 Our contributions

• Zingale & Dursi 2007:

• Under given conditions, there is a
characteristic size of a flame bubble -
larger is shredded by motions, turbulence
• This characteristic size may lead to new
flame model
 One-zone ignition
• Ignition: interaction 9e+09

of: 8e+09

7e+09

• `chemistry’ 6e+09
Ignition Delay Time

Temp (K)

•
5e+09

hydrodynamics
4e+09

• conduction 3e+09

2e+09

• Even in one-zone 1e+09

0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04 0.045
model (no hydro, Time (s)

condn) not trivial

 One-zone ignition
• `map out’ ignition times 10.0

for various conditions 9.8

•
1 picose
c
Density, temp, comp. 9.6

log10(temp)
9.4 1 nanosec

• Measurement of a 9.2 1 microsec

`burning time’ 1 millisec

9.0

1 sec

• ignition if burning time 8.8 million sec 1000 sec

faster than flow, 8.6

8.0 8.5 9.0 9.5
conduction times log10(dens)
 One-zone ignition
• ~7000 calculations
1010

• Can build ignition time

105
fitting function

Fit Ignition Time (s)

•
100

Quite good over ~15

orders of magnitude in 10-5

ignition time
10-10

• Surprising dependence on 10-15

composition - eg, even 10-15 10-10 10-5 100
Computed Ignition Time (s)
105 1010

modest amount of Ne
One-zone ignition time
• Necessary input into
ignition models 1.2
Pred. Width

1.0

• Ignition time τi must be

compared to flow ( τh )

Pres, Energy Release

0.8

and conduction (τc ) 0.6

time scales
0.4

• Also directly gives 0.2

detonation thickness - 0.0

distance behind shock -0.001 0.000 0.001
Position behind shock (cm)
0.002 0.003

where burning occurs

 Flame ignition

• Only a small part of the

question we really want
answer to:

• Given a particular
turbulent hotspot, can
that point ignite before
eddies/conduction diffuse
it away?
Spark-ignited counterflow flames:
Mastorakos, Marchione, Ahmed, Balachandran, Cambridge
 Flame ignition
• Simplified setup:

• Spherical gaussian
hotspot, quiescent flow

• Even still, huge parameter

space

• Work with
undergraduate students
Doucette, Hiratsuka,
ongoing
 Detonation Ignition
• Ignition at a point -- cold fuel
drives hot shock
shock

• Shock must slow down to

detonation speed for
detonation to ignite ignition
pt

• How big a region?

Naively, about a shock ash
thickness (Dτi)

• About a factor of 5000

too small!
 Detonation Ignition
• Curvature strongly
modifies detonation

• Speed drops with

Detonation Speed (107 km/s)

curvature

• Beyond certain point, no

steady detonation.

• Size of region must be

~5000 detonation
thicknesses Detonation Curvature (10km)-1
 Detonation Ignition
• Corresponding physical
size of region grows
rapidly with decreasing
density

• By 5x107, already ~km for

low carbon fraction

• Hard for purely local

process to ignite
detonation

• But very easy to

spuriously numerically
ignite detonation
 Outline
• Some notes on ignition
• One-zone
• Flame (ongoing)
• Detonation
• Early burning: what happens next?
• Characteristic size of flame bubbles
• Flame model?
 First Burning
• Once a point
ignites, what can
happen?
 First Burning
• Once a point
ignites, what can
happen?
• Spherical flame
burning
outwards
 First Burning
• Once a point
ignites, what can
happen?
• Buoyant rise
 First Burning
• Once a point
ignites, what can
happen?
• Distortion by
turbulence
 First Burning
• Each has
characteristic
velocity
• Flame speed fixed
at given dens
• Others bubble-size
dependent
• Flame speed always
wins until grows
Size where flame speed
stops dominating
What happens then?

Robinson,
Dursi et al
(2003)
What happens then?

Robinson,
Dursi et al
(2003)
Iapichino et al, (2006)
 Buoyant rise
• Vortical motions
of bubbles own
rise tear it apart
• Unless flame
speed is faster
• Balance sets
characteristic
flame bubble size
Balance between two
sets bubble size
 Characteristic bubble
Size

• Sets initial condition

`flame bubble’ size
for large-scale
simulation (~0.5 km)
 Compared against
simulations
• Semi-analytic
calculations
comparisons of
velocities checked
with `real’ flame
code developed at
LBNL (Zingale, Bell,
Day, Rendleman)
• Note non-constant
flame properties!
• Simple semi-
analytic model for
this case: flame
speed ~ rise
speed between
30-50 flame
thicknesses in size
 Flame modeling
• Flame modeling
based on planar
picture of flame
• Simulations w/ ~50
flame thicknesses
• Turbulent, RT
corrugation
• But much larger
range of scales
Zingale et al, (2005)
`Fragmenting Bubbles’
Flame Model
• Volume V burning
outwards,
fragmenting into
characteristic
volumes
 Rapid decrease in bubble size
just where increased burning needed?
 Conclusions
• One zone ignition times - changes w/ Ne
• Hard to ignite detonations
• Flames burn out significantly (~km) before
they start to rise
• Initial conditions need roughly this res
• Found typical burning bubble size
• Rapid decrease in size - increased burning?
• Is planar flame model appropriate?
 Conclusions
• One zone ignition times - changes w/ Ne
• Hard to ignite detonations
• Flames burn out significantly (~km) before
they start to rise
• Initial conditions need roughly this res
• Found typical burning bubble size
• Rapid decrease in size - increased burning?
• Is planar flame model appropriate?
 Characteristic Bubble
Size
• Can also predict if
burns through centre
• If ignites within ~25
km will certainly
burn through centre
• No remaining pool
of fuel
Zingale et al, (2006)
 Sedov ignition of a
detonation
100.0
1.10
1.05
1.00
0.95
10.0 0.90
Shock Velocity (V0)

90 100 110 120

Shock Position (cm)

1.0

0.1
0.1 1.0 10.0 100.0 1000.0
Shock Position (cm)