School of Engineering

University of Edinburgh
Introduction to Fire Dynamics
for Structural Engineers
by Dr Guillermo Rein
Training School for Young Researchers
COST TU0904, Malta, April 2012
Textbooks
Introduction to fire Dynamics
by Dougal Drysdale, 3
rd
Edition,
Wiley 2011
Principles of Fire Behavior
by James G. Quintiere
~65
~170
The SFPE Handbook of Fire
protection Engineering, 4th
Edition, 2009
~46
Fire Service/Sprinkler
Room
Critical
Floor
Critical
Building
Critical
100%
time
Structural Integrity
P
r
o
c
e
s
s

C
o
m
p
l
e
t
i
o
n
Progressive
collapse
Untenable
conditions
Fire Safety: protect Lives, Property and Business
from Physical Parameters Affecting Fire Growth,
Torero and Rein, CRCpress
Boundary at 256s
H
e
a
t

r
e
l
e
a
s
e

r
a
t
e

(
k
W
)
Time
Discipline Boundaries
Fire
Structures
Heat Transfer
Fire &
Structures
Fire
Fire &
Structures
Failure of
structures at
550+X C
Structures
Lame Substitution of 1
st
kind
Structures
Fire &
Structures
0
300
600
900
1200
0.1 1.1 2.1 3.1
Burning Time [hr]
Fire
Lame Substitution of 2
nd
kind
Fire &
Structures
0
300
600
900
1200
0.1 1.1 2.1 3.1
Burning Time [hr]
Failure of
structures at
550+X C
Lame Substitution of 3
rd
kind
Before ignition After 5 minutes After 15 minutes
Ignition fuel exposed to heat
Material start to decompose giving off gasses:
pyrolysis
Ignition takes place when a flammable mixture of
fuel vapours is formed over the fuel surface
Pyrolysis video
Iris Chang and Frances Radford, 2011 MEng project
0 50 100 150 200
0
50
100
150
200
250
300
Classical theory (best fit)
Apparatus
AFM
Cone calorimeter
FPA
FIST
LIFT
Others apparatus with tungsten lamps heat source
Others apparatus with flame heat source

Experimental conditions
No black carbon coating or no information
Black carbon coating
Vertical sample
Controled atmsophere (18% <O
2
<30%)
Miscellaneous
Dashed area =experimental error
Time =2s
Heat flux =+12 / -2 % [35]
Time to ignition
Time to ignition
Experimental data for PMMA (polymer) from the literature. Thick samples
2
e
o ig
ig
q
T T
c k
4
t
|
|
.
|

\
|
' '

t
=

Heat flux
T
i
m
e

t
o

i
g
n
i
t
i
o
n
Flammability
Video from WPI (USA)
Effect of heat Release Rate on Flame height
http://www.youtube.com/watch?v=7B9-bZCCUxU&feature=player_embedded
Burning rate (per unit area)
p
h
q
m
A
' '
=
' '

from Quintiere, Principles of Fire Behaviour

Firepower Heat Release Rate
Heat release rate (HRR) is the power of the fire (energy
release per unit time)
A m h m h Q
c c
' '
A = A =

Heat Release Rate (kW) - evolves with time

Heat of combustion (kJ/kg-fuel) ~ constant
Burning rate (kg/s) - evolves with time
Burning rate per unit area (m
2
) ~ constant
Burning area (m
2
) - evolves with time
A
m
m
h
Q
c
' '
A

Note: the heat of reaction is negative for exothermic reaction, but in combustion this is always
the case, so we will drop the sign from the heat of combustion for the sake of simplicity
1.
2.
3.
Heat of Combustion
from Introduction to fire Dynamics, Drysdale, Wiley
*
IGNITION GROWTH MASS BURNING
Burning area
area of the fire A increasing with time
A
A
A
A m h Q
c
' '
A =

H
Burnout and travelling flames
m
H
t
out b
' '
=

near burn-out,
location running out of fuel
Recently ignited
by flame
burn-out
a)
b)
Flame Spread vs. Angle
http://www.youtube.com/watch?v=V8gcFX9jLGc
Rate of flame spread over strips of thin samples of balsa wood at different
angles of 15, 90, 15 and 0.
Test conducted by Aled Beswick BEng 2009
On a uniform layer of fuel ignited, spread is circular
( )
2 2
2
2
constant S if
rate spread flame
t S m h A m h Q
St R A
St R
S
dt
dR
c c
' '
A =
' '
A =
= =
= =
= =

t
t t
S
2 2 2
t t S m h Q
c
o t =
' '
A =

R
Flame spread
if flame spread is ~constant, the fire grows as t
2
A
~material properties
Tabulated fire-growths of different fire types
tsquare growth fires
8
6
4
2
0
0 240 480 720 960
slow
medium
fast
ultra-
fast
2
t Q o =

time (s)
H
R
R

(
M
W
)
Sofa fire
Peak HRR= 3 MW
Average HRR ~1 MW
residual burning
+ smouldering
from NIST http://fire.nist.gov/fire/fires
growth burn-
out
Fire Test at BRE commissioned by Arup 2009
4x4x2.4m small premise in shopping mall
190s
285s
316s
0
500
1000
1500
2000
2500
3000
3500
4000
4500
5000
0 100 200 300 400 500 600 700
H
R
R

(
k
W
)
Time (seconds)
Fire Test at BRE commissioned by Arup 2009
4x4x2.4m small premise in shopping mall
Suppression
with water
Free burning vs. Confined burning
Time (s)
B
u
r
n
i
n
g

r
a
t
e

(
g
/
m
2
.
s
)
Experimental data from slab of PMMA
(0.76m x 0.76m) at unconfined and
confined conditions
confined free burning
Smoke and walls radiate downwards to fuel items in the
compartments
p
h
q
m
A
' '
=
' '

What is flashover?
Sudden period of very rapid growth caused by
generalized ignition of fuel items in the room.
Some indicators:
Average smoke temperature of ~500-600 C
Heat flux ~20 kW/m
2
at floor level
Flames out of openings (ventilation controlled)
NOTE: These three are not definitions but indicators only
Sudden and generalized ignition
(flashover)
Mechanism for flashover:
Fire produces a plume of hot smoke
Hot smoke layer accumulates under the ceiling
Hot smoke and heated surfaces radiate downwards
Flame spread rate and rate of secondary ignition increases
Rate of burning increases
Firepower larger and smoke hotter
Feedback
loop
Flashover
Compartment fires
(a) growth period
(b) fully developed fire
(c) decay period
(a)
(b)
(c)
H
e
a
t

r
e
l
e
a
s
e

r
a
t
e

(
k
W
)
Time
Fire development in a compartment - rate of heat release as a function of time
flashover
fo
Q

max
Q

Discipline Boundaries
Fire
Structures
Heat Transfer
Fire &
Structures
GI GO
If the input is incomplete/flawed, the
subsequent analysis is flawed and cannot
be trusted for design
Fire is the input (boundary condition) to
subsequent structures analysis
Design Fires
The Titanic complied with all codes.
Lawyers can make any device legal,
only engineers can make them safe"
Prof VM Brannigan
University of Maryland
Traditional Design Fires
0
200
400
600
800
1000
1200
1400
0 30 60 90 120 150 180 210 240
Time (minutes)
T
e
m
p
e
r
a
t
u
r
e

(

C
)
EC - Short
EC - Long
Standard
Standard Fire ~1917
Swedish Curves ~1972
Eurocode Parametric Curve ~1995
Traditional methods are based on experiments
conducted in small compartment experiments
(~3 m
3
)
1. Traditional methods assume uniformfires that lead
to uniform fire temperatures (?)
2. Traditional methods have been said to be
conservative (?)
Stern-Gottfried et al, Fire Risk Management 2009
Traditional Methods
Limitations
For example, limitations according Eurocode:
Near rectangular enclosures
Floor areas < 500 m
2
Heights < 4 m
No ceilings openings
Only medium thermalinertia lining
Proposed WTC Transit Hub
< 500 m2 floor?
<4 m high?
Excel, London
Rectangular?
No ceiling opening?
Arup Campus
Arup/Peter Cook/VIEW

Shard
Renzo Piano
Insulating lining?
Edinburgh Survey 3,080 compartments
18501990 buildings: 66%of volume within limitations
2008 building: 8%
Modern architecture increasingly produces buildings out of range
Jonsdottir et al
Fire Risk Management 2009
Traditional methods are based on experiments
conducted in small compartment experiments
(~3 m
3
)
1. Traditional methods assume uniformfires that lead
to uniform fire temperatures (?)
2. Traditional methods have been said to be
conservative (?)
Stern-Gottfried et al, Fire Risk Management 2009
Traditional Methods
Fuel Load
Mixed livingroom/ office space
Fuel load is ~ 32 kg/ m
2
Set-up Design for robustness and high repeatability
Compartment Temperature
Stern-Gottfried et al., Fire Safety Journal 45, pp. 249261, 2010. doi:10.1016/ j.firesaf.2010.03.007
Cardington Results
Peak local temperatures range from 23% to 75% above
compartment average, with a mean of 38%
Local minimum temperatures range from 29% to 99%
below compartment average, with a mean of 49%
Temperature Distributions
Travelling Fires
Real fires have been observed to travel
WTC Towers 2001
Torre Windsor 2005
Delft Faculty 2008
Experimental data indicate fires travel
in large compartments
In larger compartments, the fire does
not burn uniformly but burns locally
and spreads
Design Fires
Problems cannot be solved by the
level of awareness that created
them"
Attributed to A Einstein
Fire environment split
into two:
Near-field 1000-1200 C
Far-field 200-1200 C
(Alpers correlation)
T
e
m
p
e
r
a
t
u
r
e
Distance
Total burning
duration is a function
of the area of the fire
Travelling Fires
Each structural element sees a combination
of Near Field and Far Field temperatures
as the fire travels
Travelling Fires
Stern-Gottfried et al, SPFE PBD, 2010, Lund
Example 25% Floor Area fire in a 1000 m
2
Near field temperature 1200C for 19 min
Far field temperature ~ 800C for 76 min
Structural
Element
Core
0
200
400
600
800
1000
1200
1400
0 50 100 150 200 250 300 350 400
Time (min)
T
e
m
p
e
r
a
t
u
r
e

(

C
)
Point B, Rebar temperature
Point B, Gas temperature
Structural Results Rebar Temperature
0
100
200
300
400
500
0,1 1 10 100
R
e
b
a
r

T
e
m
p
e
r
a
t
u
r
e

(

C
)
Time (hours)
2.5%
5%
10%
25%
50%
100%
Case Study:
Generic MultiStorey Concrete Structure
Law et al, Engineering Structures2011
Stern-Gottfried et al, SPFE PBD, 2010, Lund
400C
0C
600 minutes
1200 minutes
T
e
m
p
e
r
a
t
u
r
e
Time
100% burn area
Rebar Temperature
Using a 3D Finite Element Model
400C
0C
600 minutes
1200 minutes
T
e
m
p
e
r
a
t
u
r
e
Time
50% burn area
100% burn area
Rebar Temperature
400C
0C
600 minutes
1200 minutes
T
e
m
p
e
r
a
t
u
r
e
Time
50% burn area
25% burn area
100% burn area
Rebar Temperature
400C
0C
600 minutes
1200 minutes
T
e
m
p
e
r
a
t
u
r
e
Time
50% burn area
10% burn area
25% burn area
100% burn area
Rebar Temperature
400C
0C
600 minutes
1200 minutes
T
e
m
p
e
r
a
t
u
r
e
Time
50% burn area
5% burn area
10% burn area
25% burn area
100% burn area
Rebar Temperature
50% burn area
400C
0C
600 minutes
1200 minutes
T
e
m
p
e
r
a
t
u
r
e
Time
2.5% burn area
5% burn area
10% burn area
25% burn area
100% burn area
Rebar Temperature
Law et al, Engineering Structures2011
0
100
200
300
400
500
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
R
e
b
a
r

T
e
m
p
e
r
a
t
u
r
e

(

C
)
Burning Area
Travelling Fires
Standard Fire - 1h 18min
EC Short
EC Long
Max Rebar Temperatures vs. Fire Size
1h 18 min
Law et al, Engineering Structures2011
0
0,05
0,1
0,15
0,2
0,25
0,3
0,35
0% 20% 40% 60% 80% 100%
D
e
f
l
e
c
t
i
o
n


(
m
)
Burning Area
Travelling Fires
Standard Fire - 1h 54min
EC Short
EC Long
Max Deflection vs. Fire Size
1h 54 min
Law et al, Engineering Structures2011
Conclusions
In large compartments, a post flashover fire
is not likely to occur, but a travelling fire
Provides range of possible fire dynamics
Novel framework complementing
traditional methods
Travelling fires give more onerous conditions
for the structure
Strengthens collaboration between fire and
structural fire engineers
Thanks
Collaborators:
J Stern-Gottfried
A Law
A Jonsdottir
M Gillie
J Torero
Sponsors:
ARUP
Jonsdottir et al, Interflam2010, Nottingham
Law et al, Engineering Structures2011
Rein et al, Interflam2007, London
Stern-Gottfried et al, SPFE PBD, 2010, Lund
Stern-Gottfried et al, Fire Risk Management 2009
Jonsdottir et al, Fire Risk Management 2009
Strengthening the bridges
Temperature of the plume
f
r
o
m

I
n
t
r
o
d
u
c
t
i
o
n

t
o

f
i
r
e

D
y
n
a
m
i
c
s
,

D
r
y
s
d
a
l
e
,

W
i
l
e
y
Burning at average heat release per unit area
where t
b
is the burning time, m is the fuel load density (kg/m
2
),
AH
c
is the effective heat of combustion and Q is the heat release
rate per unit area (MW/m
2
)
Q
h m
t
c
b

' '
A
' '
=
50 MW fire on 200 m
2
burns for 30min
50 MW fire on 1000 m
2
burns for 15min
Conservation of Mass burning time
Rein et al, Interflam2007, London
Aftermath
Average Compartment Temperature
Three different beams used
Unprotected steel I-beam
Protected steel I-beam to 60 min (12mm
high density perlite)
Concrete beam with 60 min rating
Example: Cardington
2
Results for Insulated Steel:
Parametric vs. Travelling fires
Jonsdottir et al, Interflam 2010, Nottingham
Compared to parametric fire, 110% higher temperatures
for a protected steel with 39 mm-gypsum
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0% 50% 100%
Fire area
N
o
r
m
a
l
i
z
e
d

s
t
r
e
s
s
_
Rebar temperature
Standard Fire
Parametric - Short hot
Parametric - Long cool
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0% 50% 100%
Fire area
N
o
r
m
a
l
i
z
e
d

d
e
f
l
e
c
t
i
o
n
_
Deflection
Standard Fire
Parametric - Short hot
Parametric - Long cool
0
0.01
0.02
0.03
0.04
0.05
0.06
0% 50% 100%
Fire area
N
o
r
m
a
l
i
z
e
d

s
t
r
a
i
n
_
Sagging strain
Standard Fire
Parametric - Short hot
Parametric - Long cool
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0.18
0.2
0% 50% 100%
Fire area
N
o
r
m
a
l
i
z
e
d

s
t
r
a
i
n
_
Hogging strain
Standard Fire
Parametric - Short hot
Parametric - Long cool
Structural Behaviour
Fire Progression
Core Core
Sudden Gradual
1
st
burn region 2
nd
burn region 3
rd
burn region 4
th
burn region
Base case Corner Ring - inwards Ring - outwards
Fire Shape/Path
Core
Linear
Core
Corners
Core
Ring - Inwards
Core
Ring - Outwards
Far Field Temperature Discretization
0
200
400
600
800
1000
1200
1400
0 5 10 15 20 25 30 35 40 45
T
e
m
p
e
r
a
t
u
r
e

(

C
)
Distance from End (m)
One Far Field
Two Far Fields
Monotonic Far Field
Sensitivity Results
0
100
200
300
400
500
Base case Two Far Fields Monotonic -
Sudden
Monotonic -
Gradual
Corner Ring - Inwards Ring - Outwards
T
e
m
p
e
r
a
t
u
r
e

(

C
)
0
0,05
0,1
0,15
0,2
0,25
0,3
0,35
Base case Two Far Fields Monotonic -
Sudden
Monotonic -
Gradual
Corner Ring - Inwards Ring - Outwards
D
e
f
l
e
c
t
i
o
n

(
m
)
Unprotected steel up to 10% higher steel temperature
(independent of fire size)
Protected steel from 65%-95% higher steel
temperature
Maximum over prediction (110%) at fire areas of 5-
10%
Maximum under prediction (20%) at fire areas over
85%
The above methodology was applied to a real building, The
Informatics Forum Building of the University of
Edinburgh
T
max
-method / T
max
-parametric curve - for unprotected steel:
1,0
1,2
1,4
1,6
1,8
2,0
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
T
m
a
x
-
m
e
t
h
o
d
/
T
m
a
x
-
p
a
r
a
m
e
t
r
i
c

c
u
r
v
e
percentages of floor area
HE-A 600
HE-A 300
HE-A 200
Heron Tower
46 Storey Office
Building in City of
London
3-storey atriums
forming villages
First ever project to
consider the
robustness of a
structure in a multi-
storey fire.
ARUP
Heron Tower
ARUP
Sudden and generalized ignition
(flashover)
When feedback heat flux is ~20 kW/m
2
(above the critical
ignition for most known fuels) enhanced flame spread and fast
secondary ignition take places in the compartment onset of
flashover
4
~ T q o
' '
EM-DAT International Disaster Database, Universit catholique de Louvain, Belgium. www.emdat.be
Explosions and Fire
,
,
,
,
,
,
,
,
,
,
NOTE: Immediate fatalities as a proxy to overall damage. Disaster defined as >10 fatalities, >100
people affected, state of emergency or call for international assistance.
Jocelyn Hofman, Fire Safety Engineering in Coal Mines MSc Dissertation, University of Edinburgh, 2010
Technological Disasters 19002000
,
,
,
,
EM-DAT International Disaster Database, Universit catholique de Louvain, Belgium. www.emdat.be
Jocelyn Hofman, Fire Safety Engineering in Coal Mines MSc Dissertation, University of Edinburgh, 2010
Technological Disasters 19002000
Fire and Explosions
Candle burning on Earth (1g) and
in microgravity inside the ISS (~0g)
Buoyancy
Family of possible fires
0
200
400
600
800
1000
1200
1400
0,1 1 10 100
F
a
r

F
i
e
l
d

T
e
m
p
e
r
a
t
u
r
e

(

C
)
Time (hours)
1%
2.5%
5%
10%
25%
50%
100%
Std Fire
EC Short
EC Long
Stern-Gottfried et al, SPFE PBD, 2010, Lund
Far Field Temperature
Maximum temperature at ceiling jet. Average
calculated over the correlation with the distance
from the fire (Alperts correlation)
0
200
400
600
800
1000
1200
1400
0 10 20 30 40 50 60
C
e
i
l
i
n
g

J
e
t

T
e
m
p
e
r
a
t
u
r
e

(

C
)
Distance from Fire (m)
( )
H
r Q 38 . 5
T T
3 2
max

=

nf ff
r
r
4
4
ff
r r
dr T
T
ff
nf
max

=
}
Mass flow of combustion products at the flame:
(Atmospheric air is 21% Oxygen, MW
air
=29 g/mol)
Products of Combustion
( ) 16 1 ~
21 . 0
4
1
,
+
|
|
.
|

\
|
+
+ = + = m
y x
MW
MW
m m m m
fuel
air
air st pc

ent ent pc smoke
m m m m

~ + =
Flow of products
of combustion
Smoke is mostly made of entrained air
Most of the smoke is N
2
!
>>
pc ent
m m

fuel flow rate
by pyrolysis
flow of
stoichiometric air
eg, value for propane
Ventilation flows
Mass flow of air into compartment (kg/s)
Opening area (m
2
)
Height of opening (m)
o
o
a
H
A
m

0 0 0
H A m vA m =

0
2
gH v A =
0 0 max ,
5 . 0 H A m
a
=

a a
m m

>
max ,
ventilation factor
o o
A H ,
The flow through openings has a
maximum possible limit.
At steady state, flow of smoke out
is approximately equal to the flow
of air in.
for buoyant flows
Flows in and out of the compartment are controlled by
buoyancy which scales with the density differences and
the size of the opening.
Pyrolysis of a solid
Pyrolysis
When a solid material heats up, it eventually reaches a temperature threshold
where it begins to chemically break down. This process is called pyrolysis and is
similar to gasification but with one key difference pyrolysis is the simultaneous
change of chemical composition (eg, long hydrocarbon chains to shorter chains)
and physical phase (ie, solid or liquid to vapour) and is irreversible. When a solid
is burning with a flame, it is actually the pyrolysis vapours (aka pyrolyzate)
directly above it that is burning, not the solid itself.
Pyrolysis of a liquid
Flame spread is inversely
proportional to the time to
ignition
Flame Spread rate of area growth
s
o
ig
s
t
S
o

Downward Upward
2
4
|
|
.
|

\
|
' '

=
e
o ig
ig
q
T T
c k t

t
Ignition fuel exposed to heat
Flammable mixture
T
(t3)
T
(t2)
T
(t ignition)
T
(t1)
T
ambient (t0)
Heat flux
Pilot
Temperature
D
e
p
t
h
time
Material start to decompose giving off gasses:
pyrolysis
Ignition takes place when a flammable mixture of
fuel vapours is formed over the fuel surface
Flame Spread vs. Angle
Upward spread up to 20 times faster than downward spread
upward
vertical
spread
downward
vertical
spread
Examples of HRR
workstation
mattress
wood crib
0:00 min 4:15 min 5:00 min
Polypropylene: burning inside a small compartment (0.4m cube)
Under Ventilated fires and
External flaming
Ceiling Jet
from Alpert, Ceiling jet flows, SFPE handbook
Size Matters
Surface Area to Volume Ratio vs Floor Area for a 3m High Square Compartment
0
0.5
1
1.5
2
2.5
3
0 500 1000 1500 2000 2500 3000
Floor Area (m)
S
u
r
f
a
c
e

A
r
e
a
/
V
o
l
u
m
e

(
1
/
m
)
Fire Tests
Real Buildings
Stern-Gottfried et al, Fire Risk Management 2009
Encouraging initial reactions to this work
Abstract submitted in 2007 to Structures in Fire (SiF)
Title: ON THE STRUCTURAL DESIGN FIRES FOR
VERY LARGE ENCLOSURES
Reviewer #1: This abstract does not it fit with [conference] theme.
Reviewer #2: This paper is outside the scope of the conference
Reviewer #3: The authors are encouraged to submit their paper
somewhere else
Abstract submitted in 2011 to Structures in Fire (SiF)
Title: TRAVELLING FIRES IN LARGE
COMPARTMENTS: MOST SEVERE POSSIBLE
SCENARIOS FOR STRUCTURAL DESIGN
Reviewer 1: Several works has been done and published
Reviewer 2: No significant input
Reviewer 3: Authors must provide examples for typical case studies
Thanks