Report Mo.

FAA-RD-70-72

CRASH FIRE HAZARD EVALUATION

OF JET FUELS

Andrew J. Atkinson
Thor I. Eklund
Naval Air Propulsion Test Center /
Aeronautical Engine Department
Naval Base, Philadelphia, Pa. 19112

.. ~

JANUARY 1971

FINAL REPORT
Availability is unlimited.· Document may be released to the
National Technical Information Service, Springfield, Virginia
22151, for sale to the public.

Prepared for

FEDERAL AVIATION ADMINISTRATION

Aircraft Development Service
Washington D. C., 20590
Reproduced by
NATIONAL TECHNICAL
INFORMATION SERVICE
Springfield, Va. 22151
 The contents of this report reflect the views of the National
Aviation Facilities Experimental Center, Atlantic City, New
Jersey, which is responsible for the facts and the accuracy of the
data presented herein. The contents do not necessarily reflect the
official views or policy of the Department of Transportation. This
report does not constitute a standard, specification or regulation.

The Federal Aviation Administration is responsible for the promotion, regulation

and safety of civil aviation and for the development and operation of a common
system of air navigation and air traffic control facilities which provides for the
safe and efficient use of airspace by both civil and military aircraft•

. !The National Aviation Facilities Experimental Center maintains laboratories,

facilities, skills and services to support FAA research, development and imple-
mentation programs through analysis, experimentation and evaluation of aviation
concepts, procedures, systems and equipment.
 TECHNICAL REPORT STANDARD TITLE PAGE
~--::---~-.----------,--=--=----..,----~-----~~--....,......"....-..,.-~------_."
1. Repo.t No. 2. Govern",ent Acce ..len No. 3. Recipient's Catalog No.

FAA-RD-7o-72 i
! 4. T I tie and Subt; tl. S. Repol! O n - t - e - - - - - - - - · - - - - - <

l January 1971
j 6. p •• fonning Orgoniza'ion W;-------_._-_.
1 CRASH FIRE HAZARD EVALUATION OF JET FUELS i
1"" ."thorl sl B. P.rforming Orgoni lotion Report No,
i Andrew J. Atkinson
iI Thor I. Eklund FAA-NA-70-64
, 9. Perlnrming Orgonl\"otlon Nom. ond Addr... 10. Work Unit No.
I NAVal Air Propulsion Test Center Project No. 520-005-0J+X
i Aeronautical Engine Department II. Controct or Grant No.

i Naval Base, Philadelphia, Pennsylvania 19112 ~~D~O.J FA66NF-AP-14

13_ Typ't of Report and Period- Cov;.;;;-----
.i . ------------------------------------01
12_ Spo",oring Agency Name and Addr...

Federal Aviation AdmL1istration

Systems Research and Development Service k-=-...lF~ii.n
n1aA1"U"i.epQ1j:. . _
14. Sponsoring Agency Code
Washington, D.C. 20590
__ - ... 1- _
I....~-
I : S. Supp 'em"rotary Note.

Ii
r-------
16, Abstract

! An investigation was conducted to determine the relative crash f'ire hazards of

jet fuels under survivable crash conditions. Kerosene, JP-4, and mixtures of
both tolere evaluated under various release modes (pools, drips, streams and
sprays) and in the presence of possi.ble ignition sources (electrical sparks,
friction sparks, open flames, and hot surfaces). Wind speed, wind air t!=,mpel'8ture
end fl'cl temperatures were also 'faried> The r~sults of this evaluation land t.tJe
conclusions reached are discussed in the report.

,
.
.
_-_._----_._-~--

j(~y Words 18. Distlibution Statement

Crash Fire Safety Availability is unlimited. Document may
Turbine Fuels be released to the Clearinghou se for
Ignitability Federal Scientific and Technic 81 Informa-
Combustion tion, Springfield, Virginia 22 151, for
Flame Propagation sale to the public.
Security Classil. (of this r.port) 20. S.curity Clauil. (of this page) 21. No. of Pages
----------

Unclassified Unclassified 54
Form DOT F 1700.7 (1-691
, I
 PREFACE
This report was prepared by the Naval Air Propulsion Test .
Center, Aeronautical Engine Department, Naval Base,
Philadelphia, Pennsylvania, for the Federal Aviation
Administration. The work was part of a program of the
Aircraft Division, Systems Research and Development Service,
Washington, D.C. The work was administered under the
direction of Mr. Samuel V. Zinn, Jr., who served as Project
Manager for the Instruments and Flight Test Section,
Aircraft Branch, Test and Evaluation Division, National
Aviation Facilities Experimental Center, Atlantic City,
New Jersey.

iii
 TABLE OF CONTENTS

INTRODUCTION 1

TEST PRffiRAM 1
DISCUSSION 2

Dripping Mode 2

Stream. Mode 3
Pool Mode 9
Spray Mode 22

CONCLUSIONS 33

APPENDICES
A Description of the Test Equipment 1-1

B Properties of the Fuels Tested 2-1

C References 3...1

Preceding page blank

v
 LIST OF ILLUSTRATIONS

FIGURE Page
1 The Effect of Wind Velocity on the Minimum Fuel
Temperature Required for Flame Climbing up a
Fuel Stream 4

2 The Effect of the Spark Ignitor Location on the

Minimum Fuel Temperature Required for Ignition
and for Flame Climbing up a Fuel Stream 5
3 The Spatial Ignition Envelope with Flame Climbing
at a Fuel Temperature of 250 0 F 6
4 The Effect of Wind Velocity on the Minimum Surface
Ignition Temperature for Fuel Streams 8

5 The Effect of Liquid Fuel Temperature on the Rate

of Flame Spread across a Fuel Pool 11
6 The Effect of Initial Fuel Temperature on the
Average Wind Blowout Velocity of Burning Fuel
fools 12

7 The Effect of Pool Diameter on the Average Wind

Blowout Velocity of Burning Fuel Pools 14
8 The Effect of Spark Ignitot Height on the Minimum
Fuel Temperature Required for Ignition of Shrouded
Fuel Pools 15
9 The Effect of Open Flame Ignitor Height on the
Minimum Fuel Temperature Required for Ignition of
Shrouded Fuel Pools 16
10 The Effect of Open Flame Ignitor Height on the
Minimum Fuel Temperature Required for Ignition
of Unshrouded Fuel Pools 18
11 The Effect of Spark Ignitor Height on the Minimum
Fuel Temperature Required for Ignition of
Unshrouded Fuel Pools 19
12 The Effect of Hot Wire Ignitor Height on the
Minimum Fuel Temperature Required for Ignition
of Unshrouded Fuel Pools 21

Preceding page blank

vii
 LIST OF ILLUSTRATIONS (Con't.)

FIGURE Page

13 The Effect of Wind Velocity on the Minimum Fuel

Temperature Required to Ignite a Fuel Pool 23
The Effect of Fuel Blend Composition on Flash
Point and on the Average Minimum Temperature
at Ignition (Open Flame Ignition Source Placed
One Inch above Pool Center) 24-

15 The Effect of Fuel Spray Temperature on the

Minimum Ignition Energy Requirements 26
16 The Effect of Fuel-Air Ratio on Flame Propagation
Rate through Fuel Sprays 28
17 The Effect of Fuel Temperature on the Flame
Propagation Rate through Fuel Sprays 30
18 The Effect of Wind Velocity on the Minimum
Surface Ignition Temperature for Fuel Sprays 32

APPENDIX A

1-1 Schematic of Wind Tunnel Test Section 1-2

1-2 Wind Tunnel Test Apparatus 1-3

1-3 Schematic of the Test Section Configuration for

Stream Ignition and Flame Climbing Tests

1-4 Burning Stream after Ignition and During Flame

Climbing 1-5
Schematic of the Test Section for the Hot Surface
Ignition Tests
Hot Surface Ignition Apparatus

viii
 LIST OF ILLUSTRATIONS (Con't.)
FIGURE Page
1-7 Schematic of the Test Section Configurations
for Pool Tests 1-8

1-8 Typical Pool Fire 1-9

1-9 Schematic of the Apparatus Employed for the Flame
Speed Tests JrlO
1-10 Flame Speed Test Apparatus 1-11

ix
 LIST OF TABLES

TABLE
1 The Minimum Fuel Temperature for the Ignition
of JP-4 and Kerosene Using a Friction Spark
Ignition Source 20

APPENDIX B

2-1 Typical Properties of the Test Fuels 2-1

Preceding page blank

xi
 INTRODUCTION

The analyses of survivable aircraft crashes (where the impact force

is not of a fatal level) indicate that post-crash fires can and do
seriously increase fatality rates (1). In an effort to define, eliminate,
or mitigate the causes of this hazard various investigations have been
conducted. Pinkel, et al.(2, 3) have reported that a major cause of post-
crash fire is the ignition of fuel sprays by hot surfaces with both piston
engine and jet engine aircraft. Scull (4) has made a literature search
on the flammability of hydrocarbons under various conditions, but the
majority of the data were not obtained with actual jet fuels. In an
attempt to eliminate the fire hazard, various approaches have been studied;
these include the modification of fuel tanks, the elimination of ignition
sources, and the alteration of fuel characteristics (5). There has not,
however, been any systematic test program conducted to specifically eval-
uate the relative safety hazards of JP-4 and aviation kerosene.

The Federal Aviation Administration initiated this program to conduct

such an evaluation. The purpose was to determine the relative safety
hazards of JP-4, aviation kerosene, and mixtures of both under typical
environmental conditions. The aviation kerosene used in this program was
the Navy fuel, JP-5. The civilian counterpart for this fuel is Jet A,
also an aviation kerosene. Since the flash point and specific gravity of
Jet A are lower than those of JP-5, the data curves for Jet A where vol-
atility is a factor would show slightly less difference between kerosene
and JP-4 than is shown in this report. The fuels were evaluated under
various release modes (pools, drips, streams, and sprays) in the presence
of possible ignition sources (electric sparks, friction sparks, open
flames, and hot surfaces) and with variations in wind speed, wind air
temperature, and fuel temperature. The project was divided into two phases.
In the first phase a literature search was conducted in order to avoid
duplicating valid and applicable work done by others. The second phase
focused on determining the effect of those parameters influencing fuel
flammability not covered or inadequately covered in the surveyed literature,
and designing suitable test apparatus to achieve this end.

This report presents the details of the work performed, the data
obtained, and the conclusions established.

TEST PROORAM

The test program was established to conduct a hazard evaluation of

aviation fuels under the various release modes likely in an aircraft
crash (drips, streams, pools, and sprays), each in the presence of possi-
bIB ignition sources (open flames, hot surfaces, electrical sparks, and
friction sparks). Also included in the program were variations in wind
velocity, wind air temperature, and fuel temperature. Details of the
test procedures will be given in the discussion section of the report.

1
 DISCUSS ION

Dripping ~

Testing was inib.ated with the open flame Ignition of fuel drops.
The illvestigation ws,s aimed at finding those conditlOns under which a
flame would climb a column of dripping fueL 'fne evaluation of this
phenomenon is important since it determines the ability of a flame to
prcpagate to the fuel source and potentially cause a catastrophe. The
test, consisted of dripping fuel from a needle t~p, at a predetermined
rate, into a pan of burning fueL Th1.S pan was posit loned approximately
twenty inches below the fuel source 0 Twenty inches was the maximum .
distance the fuel source could be placed above the open flame. The
ll18ximum distance was chJsen for the tests since at that position the
open flame exerts the least aerodynamic and heating effect on the flow
fiEld immediately ar~:>und the fuel source, Tbe temperature of' the fuel
was increased until the flame propagated up the column of dripping
fuel to the fuel source. Drop size was varied from 0"20 to 0 25 0

centimeters 9lld fuel flow rate ranged from zero to that rate which
produced a solid stream, With the use of needle tiPS E>8ch drop size
remained constant over the entire fuel flow range. At each drop size
and flow rate condition fuel temperature was inc"'e8s~d from ambient
(65 C F) to 175°F for JP~4 and 250°F for kerosene, Flame climbing was not
observed with either JP=4 or aviatio'1 kerosene under the range of con-
ditions tested. The climbing phenomenon wa.s observed only with a.
steady stream of fueL

As two consecutive drops fall from the fuel son!'C'e. the distance
between them increases with time. There is obvi)usly some minimum
distance at which a burning drop can ignite the drop immediately above
it. This minimum distance is dependent on the drop size. the flame
height to drop diameter ratio) the ignition lag tim~" and the environ=
mental conditions. In addH Lm to t.he testing discussed;, a t,heoretical
study was conducted ass 1l!!ong the more fav,)rable i.gni tion 8ond~tians"
Consecutive drops were assumed t'J be touChlllg as they left the fuel
source, a flame height to drop diamet.:or raba ~f 10 was empl::lyedm the
calculations (6, 7)J and ignition delay time of 0 56S, was assumed.
Calculations were done for drop diameters of 0,20 em and 0,,60 em. These
calculations indicate f1.8me cl.imbmg vnll not oc.:u~ unless the ignition
source is approxll!lBtely 4 inches from the fuel t:ourrce. S.inee deviations
from these ideal ?.:ondit~ons are a certainty) it is considered that the
occ~rence of a flame;limbing a drippwg column Df fuel is extrerrely
remote.

2
stream M2.@
This investigation as with the dripping mode evaluation~ was
j

centered on finding those conditions under which a flame would climb

up a colunm of fuel to the fuel source. The initial tests consisted
of ejecting fuel from a needle tip; at a predetermined rate, into a
pan of burning fuel and determining the minimum fuel temperature for
flame climbing. The distance from the fuel source to the ignition
source was twenty inches. The stream was ejected from a 0.0.3 inch
diameter needle tip. This diameter was chosen since it was the mini~
mum diameter at which the stream remained intact, that lS~ no misting
or breakup was observed. For each test, fuel flow rate (100 ~c/min
and 200 cC/min) and wind air temperature (00F.1 85 0F and 125 0F) were
fixed. Wind velocity was varied in increments from 0 to 1.5 mph and
the minimum fuel temperature for flame e~Lmbing was noted. (Since
stream breakup occurred at wind velocities above 1.5 mph) this vel~
oeity was not exceeded.) The data obtained are plotted on figure 1
and illustrate the effect of wind velocity on the minimum fuel
temperature for flame climbing f·Jr both kerosene and JP~4. Slnce
the wind air temperature variation produced nQ discernible effect on
the JP~4 fuel temperature necessary for flame climbing) ~his variable
was omitted in the kerosene tests. The results of these tests indi~
cate that' (1) as wind velocity increases the mJ:limum fuel temperature
for flame climbing increases; (2) there is no significant effect of
fuel flow rate or wind temperature on the fuel temperature necessary
for flame climbing; and (3) the minimum temperature for flame climbing
is signifi.cantly higher for kerosene than for JP~4 under all conditions.

The second series of tests conducted with fuel streams utilized a 10

joule DC spark as the ignition source. The spark was pla~ed at various
radial locations relative to the fuel stream and at incremental vertical
distances from the stream SOl..lrce. The fuel temperature waE' varied and
the minimum fuel temperature for flame climbing was noted. The data
obtained are illustrated in fi.gure2. As the ignitor is moved away from
the stream in a radial direction) the mln~~~ tempe~ature required for
flame climbing increases significantly. At anyone radial location~ as
the spark ignitor is moved vertically away from the fuel sC1urce J the
temperature required for flame climbing decreases. For temperatures
below 225°F for kerosene and below 50'JF for ,jP~4. t,he flame travelled
downward from the ignition point but flame c.Li.mbmg did not occur In
0

the open flame ignition tests. flame climbing to the sour.::e was also
limited by these same respective temperature values.
At all ignitor positions) the minimum fuel t,emperatures for both
ignition and ignition with flame climbing were Significantly higher for
kerosene then for JP~4. For anyone fuel temperature 9 the spatial
ignition envelope for JP~4 is wider than that for kerosene. Figure.3
0
shows this comparison for a fuel temperature of .250 F 9 in which flame
climbing' occurs within the ignition envelope. For kerosene the spatial
ignition envelope ceases to ex~st at fuel temperatures below 155°F; for
JP~4 this limitation,decreases to ~.300F •

.3
 350

300
....
0
Kerosene

D)
c
.a 250 _
E
u
GI
E
g
....
... 200
...
0

... GI
;,

..
g
GI
Q.
E 150
GI
~

GI
....;,
E Symbol Wind Temp.. - OF
;,
E 100
0 0
c 0 85
125
~ '"
GI Flow Rate - cm3/min
.
D)
g
GI
0 100

~
>
50
• 200

o
o 0.4 0.8 1.2 1.6 2.0

Wind Velocity - miles per hour

FIGURE 1
The Effect of Wind Velocity on the Minimum Fuel Temperature
Required for Flame Climbing up a Fuel Stream

4
 400

""
0

I 350 Kerosene
Gl
c
:0
.-E
r; 300
II
E
J!
u.
250 -
... ~
0

. 1/

-. ::)

a
1/
a.
200

E
... 1/
150
1/
::)
u.
Symbol Vertical Dist. from
E Fuel Source- in.
::) 100
E 0 20
c 0 15
~ 10
"
1/ SO
..
m
1:1
1/
ignition; but no
flame climbing
>
c:r:
0
0 2 4 6 8

Radial Distance from Fuel Stream - in.

FIGURE 2

The Effect of the Spark Ignitor Location on the Minimum

Fuel Temperature Required for Ignition and for Flame
Climbing up a Fuel Stream

5
 FIGURE 3

The Spatial Ignition Envelope with Flame Climbing

at a Fuel Temperature of 250°F

6
 In the final series of stream tests J the hot surface ignition
characteristics of JP-4 and kerosene were investigated. 4.5 cm? of
liquid fuel were ejected onto a hot stainless steel plate 2.5 inches
below the fuel source. Wind velocity across the plate was varied and
the minimum surface ignition temperature was noted. The plate dimen-
sions were 19 inches x 26 inches with a directly heated area
9 inches x 16 inches. The fuel was ejected 6.25 inches downstream
of the leading edge of the plate (2.5 inches downstream into the
directly heated area). The plate temperatures reported are the
averages of four surface thermocouple measurements located in the
vicinity of the fuel stream impingement point. III this vicinitYJ the
maximum temperature difference between thermocouples was approXimately
80 oF.
The data obtaLYled~ as illustrated in figure 4~ show that as wind
velocity increases there is a corresponding in~rease in the surface
temperature required for ignition with both JP~4 and terosene. For
ignition to occur i both time and sufficient thermal energy (surface
temperature) are required. The higher the surface temperature 9 the
shorter the ignition delay time (the t.ime necessary for ignition).
By controlling the wind veloeitY9 the fuel vapor residence time in the
vicinity of the hot surface is also controlled. If residence time
equals or exceeds ignition delay time for the test temperature, ignition
will occur. As wind velocity is increased there is a decrease in
residence time and therefore a higher surface ignition temperature is
required.

Goodall and Ingle (8), in their tests wi th kerosene! also have

shown the effect of wind velocity on surface ignition temperature. For
wind velocities of 2 ft/sec and above? their data display the same
general linear relationship between wind velocity and surface ignition
temperature as the data obtained in this program. However? their sur--
face ignition temperatures were) for anyone wind velocity~ lOOoF lower
than those obtained in thie test program. A feasible explanation for
this difference is that their heated surface was longer than that used
in this program., thereby producing a greater residence time which caused
ignition to occur at a lower surface temperature.

Goodall and Ingle have also reported that at 2 ft/ser and below J
wind velocity has a much more pronounced effect- on surface ignition
temperature requirements. At near static conditions} their reported
surface igr.ition temperature is approximat ely 6,OoF and it rises to
1150 0 F at 2 ft/sec. This static ign1.tion temperat'~re (650°F) has also
been reported by MacDonald (9). MacDonald; s work o in this area con-
9

sisted of dropping liquid kerosene onto a heated 6 inch pipe enclosed

in an 18 inch sphere - this enclosure confines the vapor to the vicinity
of the heated surface. With the test facility used by Goodall and Ingle
the fuel vapors at the near static conditions were essentially confined
since they could not diffuse or convect more than 4 inches above the hot

7
 1450 . - - - - - - - - - - - - - - - - - - - - - - - - - _

Yo 1400
o

41

-
~

g
~ 1350
41
a.
E
41
l-
JP - 4

e
.-
-
0

1300
e
l:'l

Gl

-
u
g
~

~
\It 1250

E
~

E
.-e
~
1200

1150 -rl----r---..,r----,....---~---.,..---'""""---
o 2 3 4 5 6 7
....
.....!

Wind Velocity ft.lsec.

FIGURE 4
The Effect of Wind Velocity on the Minimum
Surface Ignition Temperature for Fuel Streams

8
surface, On the other hand~ at the higher velocities (>2 ft/sec) the
system approaches that of an open plate, since the vapors were more
readily carried away. The degree of confinement is very important.
The use of a 42 inch duct in this program permitted fuel vapors
to readily escape the heated region above the plate for all wind
velocities_ Therefore~ the time allowed for ignition is decreased.
Due to this decrease in allowable ignition time) the surface tempera-
ture required for ignition is substantially increased, and the projected
static ignition temperatures obtained are higher than the 6S0 oF reported
by other investigators. Scull (4) has reported on the ignition of fuels
dropped on heated metal plates in open air. The data indicate that the
ignition temperature of kerosene is much higher when dropped on a heated
open plate than when dropped onto a heated metal tube that confines the
mixture to a high temperature region. A comparison of these two environ-
ments for kerosene shows a surface ignition temperature of 417°F with the
confined system and 1202 0 F with an open steel plate. This latter value
agrees quite well with the projected static surface ignition temperature
obtained in this program.
The data obtained also indicate that the surface ignition tempera-
ture at anyone wind velocity is SOoF higher for JP~4 than for kerosene.
MacDonaJd (10) found the same magnitude of difference between these fuels
in his stati~ test, in which the sphere temperature enclosing the hot
surface was varied from 70 0 F to 400 oF. In the ASTM 215S-66 autoignition
temperature test for liquid petroleum products, the ignition temperature
of kerosene is again approximately SOoF higher for JP~4. For the ASTM
test, the entire environment is maintained approximately at constant
temperature.
In reviewing all of the efforts conducted, combined with the wind
velocity experiment~ in this program 9 the indication is that the ASTM
autoignition test provides a good reliable guldeLn determining the
difference in surface ignition reqUirements between fuels. The absolute
temperature requirement for each fuel will depend on w.ind velocity,
ignition surface area, surface composition and condition (4)2 fuel
.injection rate (9), and the degree to which the fuel vapors are confined
to the heated region.

One of the more comprehensive reports on the combustion process

associated with pools has been written by Glassman and Hansel (11).
Their work includes a critical literature review and remarks on such
aspects as flame spreading and steady burning of pools of flammable
liquids ~ but contains limited comment on ignitability. Their report
and their references were a prime source of information related to
pool fires.

9
 It has been concluded that for conventional fuels the most important
parameter governing the mechanisms of flame spread across a fuel surface
is the flash point of the fuel ~- or to be more precise the fire point.
Since the flash point is only a few degrees lower than the fire point,
and since it is more connnonly used as the limit.ing temperature' in safety·
applications; the flash point is used as the reference temperature in
this discussion. The flame propagation rate across a pool will be
governed by gas phase phenomena when the pool fuel temperat-ure is above
the flash point 2 and by liquid phase phenomena when the pool fuel tem-
perature is below the flash point. The reported (12) rates of flame
spread as a function of fuel temperature, for both JP~4 and aviation
kerosene~ are illustrated in figure 5. The rate of flame spread
increases" rapidly at temperatures immediately above the flash point for
each fuel, and then levels out to a connnon rate. The flame spread rate
for JP-4 is higher than that for kerosene at all temperatures below
200o~ and it is much higher at the more cammon ambient temperatures.

Blinov and Khudiahov (13) have reported on the linear burning

velocity or regression rate (the depth of fuel consumed per unit time)
as a function of pan diameter for various fuels. In addition, the
Bureau of Mines (14) has conducted tests which have provided the re-
gression rates for JP=4 and kerosene in an 8 ~ch diameter pan. A
review of the combined data indicates that there is no significant
difference in the JP-4 and kerosene linear burning rates for pan diam-
eters below 50 inches and for pan diameters above 50 inches they are
identical.
Welker (15) has published data showing the effect of wind velocity
on the angle of flame tilt for a wind blown pool fire. Various pure
compounds such as acetone, benzene, cyclohexane, n~hexane, and methanol
were used as fuels in his experiments. A derived formula and experi-
mentation demonstrate.. that the important parameters governing flame
tilt angle are the flame drag coefficient, the pan diameter, the flame
density and the ambient air density. A revIew of the foregoing work
leads to the conclusions that for a particular set of wind and tempera-
j

ture conditions) the tilt angles for JP~4 and kerosene will not differ
significantly •
Tests in this program were foeused on determining wind extinction
velocities for pool fires ~~d the comparable ignitability of pools of
JP-4, kerosene, and blends of these fuels. These areas are of obvious
interest from a fire hazard standpoint, and only limited information
could be found in the literature. Initial tests were conducted to
ascertain the wind velocity required to blow out a pool of burning fuel.
Pool diameter was varied from 6 inches to 10 inches. The range of
initial fuel temperatures tested was dependent on the fuel tyPe. All
tests were conducted at atmospheric pressure and the wind air temperature
was approximately 60°F. Figure 6 shows the data obtained with an 8 inch
pool, in which the effect of fuel type and initial fuel temperature on

10
 800 - . , - - - - - - - - - - - - - - - - - - - - - - - - -

700

c
600 Kerosene
E
".:
- 500 .
."

.aa.
GI
400 -
VI

GI
E
a 300

-..
Yo

0
200
GI

a
Ill:

100

o
-40 o 40 80 120 160 200 240 260
Liquid Fuel Temperature OF

FIGURE 5

The Effect of Liquid Fuel Temperature on the

Rate of Flame Spread Across a Fuel Pool (Reference 12)

11
 20 .. -
U
GI
til Pool Diameter = 8 inches
.....:
-I
16
>-
- u
0
GI
>
- ; :) JP - 4
0 12 o ~
f-'
I\) ~ c
0 o
aa 17 C

-g
Kerosene c
c
~
8
GI
01
a
~

GI
>
~

4 60 , i I
~;---~;----:-;;;:---:;:~----,r-----r----~---......,_J I , , , I
80 100 120 140 160 180 200

Initial Fuel Temperature - of

FIGURE 6

The Effect of Initial Fuel Temperature on the Average

Wind Blowout Velocity of Burning Fuel Pools
wind blowout velocity can be noted. Initial fuel temperature did not
affect the average wind blowout velocity, and the difference measured
between fuel types is considered insignificant. Hirst and Sutton (16)
have also reported that extinction velocity was not affected by initial
fuel temperature in their experiments with kerosene at atmospheric
pressure. However, they found that there was a temperature effect at
pressures below 0.4 atmospheres. At these lower pressures, increases
in fuel temperature effect an increase in the extinction velocity re-
quired. Figure 7 illustrates the effect of pool diameter. As the
diameter of the pool increases there is a corresponding increase in
the wind blowout velocity. The difference in values obtained with JP-4
and kerosene is not considered significant. From the blowout' velocity
tests conducted, and from the review of all applicable work, it is
expected that the extinction velocities for JP~4 and kerosene will be
approximately the same for all pool diameters.
Upon completion of the wind blowout velocity testing, experimentation
was aimed at determining the ignition characteristics of fuel pools with
various ignition sources. In the initial test a 10 joule DC spark was
placed at various heights above the center of a 10 inch pool of JP-4 and
the minimum fuel temperature for ignition was measured. It was observed,
however, that ignition occurred (even at elevated temperatures) only
when the ignition source was placed extremely close to the pool surface.
Since the fuel vapors were not diffusing more than 1 inch vertically and
therefore relative data would be difficult to obtain (it was assumed that
the kerosene height range would be even smaller), a 12 inch high shroud
was placed around the pool to circumvent this problem. A two inch wide
opening at the downstream end of the shroud allowed ignition source place-
ment and also allowed some fuel vapor to escape. The partially shrouded
pool simulates a spatial position internal to a much larger pool, in
which fuel vapor would more readily diffuse upward. The plotted results
of these tests are shown on figure 8. As the height of the ignitor above
the pool was increaeed, the minimum fuel temperature for ignition also
increased; and for all ignitor heights the minimum fuel temperature for
ignition was higher for kerosene than for JP-4.
An open flame was the second ignition source used in the pool ignition
experiments. These tests were conducted using the same pool shroud em-
ployed in the previous electrical spark tests. The data obtained are
illustrated by the solid lines on figure 9. As the height of the flame
above the pool was increased, the minimum fuel temperature required for
ignition increased correspondingly. For all ignitor heights above the
pool surface, the minimum fuel temperature required for ignition was
again higher for kerosene than for JP·-4. Also plotted on figure 9 are
the electrical spark data. Minimum fuel ignition temperatures for both
fuels are lower with open flame ignition than with electrical spark
ignition for all ignitor heights. Since the continuous flame ignition
source creates air convection currents, which can sweep the fuel pool and
carry fuel vapors to the flame., ignition is enhanced. Convection, then,
can play an important role in the ignition of these fuels.

13
 20

16 _
..;
GI
til

-..
"':

12

...
>-

-0
GI
>
..
:I
8
0
~
0
IIllI

"'0
C
4
~

GI
DI
.a
GI
>
~
0 -
6 7 8 9 10

Pool Diameter inches

FIGURE 7
The Effect of Pool Diameter on the Average
Wind Blowout Velocity of Burning Fuel Pools

14
 250 -......- - - - - - - - - - - - - - - - - - - - - - - - - - _

200 . Kerosene

Y- o
0

...-
0 150 .

c
Dl

..
"
...... -
ell
;:)
100

" ell
a. JP-4
E
... CIl

50
ell
;:)
Y-

E
;:)

E
.- 0
c
~
a -
CIl
Dl

".
ell
>
~
-50 -
a 2 3 4 5
Height of Ignitor above Pool inches

FIGURE g

The Effect of Spark Ignitor Height on the Minimum

Fuel Temperature Required for Ignition of Shrouded Fuel Pools

15
 250

..••.'
..•• ..
•• ••

.... ..
U-
0 Kerosene Spark ....
....
200 . ••
..
.. .. ......
••
....•...
....
c
0
.. ..
c
m
, .... •••••••••••••••
150 -

CJ ••
.
Ql

-.
~

CJ
G) 100
a.
E
... III
J P -4 Spark

III

U-
~

50
............... Open Flame
E
~ ••
E
••••••

~
c
....
Ql 0
m
..
CJ
III
>
c(

-50
o 2 3 4 5 6
Height of Ignitor above Pool inches

FIGURE 9

The Effect of Open Flame Ignitor Height on the Minimum Fuel

Temperature Required for Ignition of Shrouded Fuel Pools

16
 Since ignition with the open flame occurred at lower fuel tempera-
tures, 'it was decided to experiment with unshrouded pools employing the
open flame as an ignition source. Due to the forced convection provided
by the open flame, unshrouded fuel pools were able to be ignited at
various ignition heights. The data obtained, which are illustrated in
figure 10, are similar to those obtained in other ignition eXperiments. '
That is, the minimum fuel temperature for ignition increased as the
height of the ignition source was increased, and the fuel ignition
temperatures were higher for kerosene than for JP-4 for all ignitor
heights investigated.
In order to gain a better understanding of the ignition phenomena
involved, a more thorough investigation of the ignition characteristics.
of unshrouded pools was pursued with the electrical spark as the ignition
source. The results of these tests are illustrated in figure 11 along
with the open flame data. For kerosene the minimum fuel temperature for
ignition was higher when using electrical sparks than when using the open
flame; and as the ignition height increased, there was a more pronounced
increase in the fuel temperature required for ignition. In concert with
the earlier unshrouded pool tests, experiments conducted with pools of
JP-4 using the electrical spark as the ignition source yielded limited
data. At an ignitor height of one inch, ignition occurred at ~15Or and
repeatability was good. At ignitor heights above one inch but below two
inches, ignition occurred, but at such erratic temperature levels that
no reliable data point could be reported. With a spark ignitor height
of two inches ignition did not occur even with the JP-4 fuel temperature
as hi8h as 190 oF.
Visual observation of the JP-4 vapors indicate they rise to a height
of approximately one inch and then move horizontally along the test
tunnel floor. Slight, but unavoidable, drafts influence the flow of these
vapors to such an extent that they can cause random vapor flow upward to
the ignition source,and hence ignition occ'crrs. The irregularity of these
drafts causes irregular flow and hence large scatter in the minimum fuel
temperature at ignition. In contrast to the vapor flow behavior Observed
with JP-4, visual observation of the tests with kerosene clearly shows a
convection effect in which the fuel vapor~air mixture rises to the spark
ignition source (as opposed to pure diffusion). Due to the predominant
convection effect with kerosene, its ignition characteristics could be
consistently established.

The question naturally arises as to why, in the spark ignition tests,

kerosene vapors are convected upward with heated air while JP-4 vapors
are not. Convective air flow in the pool ignition experiments can be
caused in two ways in a quiescent atmosphere. First, as the fuel is
heated above ambient temperature, it can heat the air immediately above
the fuel pool surface causing it to convect upward. Second, an ignition
source such as an open flame can create convection flow around itself
which causes the air below to flow over the fuel pool and rise. In the
electrical spark tests, convection can occur only through heat transfer
from the hot fuel surface to the air above.
17
 200 .

u.
0
Kerosene
g 0

c
..
0 150 - 0

C
DI

..
Cl

.....
Gl
::J
100
Cl
Gl
11.
E
.-Gl
Gl
SO -
:J
u.

E
::J
E
c
~ 0

Gl

..
DI
Cl
Gl
>
<C
-SO
0 2 3 4 5
Height of Ignitor above Pool inches

FIGURE 10

The Effect of Open Flame Ignitor Height on the Minimum Fuel

Temperature Required for Ignition of Unshrouded Fuel Pools
~
o 200 -r------------------------~

c
..
o
'8 • • • • • • • • • • • •
.....•.•••....•.•••••....•••................•.....
150 - Kerosene - Open Flame

..o
..
Ql

o
~ 100 -
Q.
E
Ql
to-
.... ..-'
.. .. ....
....
so
....•.
.....-
E
~
.-....
E JP-4 - Open Flame
.... ... ••••
c
.... ....
.. -... ..
Ql
o •
..... ..
....o
01
c .4----J P-4 Spark
>
c(

-so -t-----r------P-----.....--------.....J
o 2 3 4 5
Height of Ignitor above Pool - inches

FIGURE 11

The Effect of Spark Ignitor Height on the Minimum Fuel

Temperature Required for Ignition of Unshrouded Fuel Pools

19
 'There is a high concentration of JP-4 fuel vapors above the fuel
pool~ at ambient temperatures and above, which rise to approximately
one inch above the pool surface. This layer of vapor, in effect,
insulates the hot fuel surface from the surrounding ambient air. There-
fore, since efficient heat transfer from the liquid fuel to the air does
not occur, sufficiently strong air convection currents are not realized'.
In addition, calculations show that for ambient fuel temperatures and
above~ the JP-4 vapor-air mixture is denser than the surrounding ambient
air. Therefore, natural convection will not occur and hence the vapor-
air mixture will not rise. Without these convection currents the possi-
bility of ignition is significantly mitigated.
Kerosene has a much lower vapor pressure than JP-~and there is
almost no vapor buildup above the pool surface. Heat transfer from the
hot pool surface to the surrounding ambient air is therefore reasonably
efficien~and convection currents do occur. In addition~ because of the
low vapor pressure of kerosene at the fuel temperaturestested~ the vapor-
air mixtures'were less dense than the 600 F - BOoF ambient air. This
mixture then~ will convect upward.

In these tests~ the lack of convection air currents significantly

reduced the chance for ignition of JP-4 fuel pools. It should be
realized however that in a real environmental situation air currents
will always exist. Consequently, JP-4 ignition at ambient temperature
can occur ~ while the limiting height will be de,termined by the flow
characteristics of these air currents. In a non-quiescent environment
then~ the difference in ignitability between JP-4 and kerosene fuel pools
can be attributed primarily to flash point differences.
The next series of unshrouded pool tests was conducted using a hot
nichrome wire as the ignition source. The data obtained are illustrated
in figure 12. Again, minimum fuel ignition temperature increased as the
height of the ignitor above the pool was increased~ and for all ignitor
heights the average minimum fuel temperature for ignition was higher for
kerosene than for JP-4.
Pool testing with friction sparks as the ignition source also indi-
cated that higher fuel temperatures are required for the ignition of
kerosene as compared to JP-4. The sparks were produced through friction
contact between a six inch diameter Simonds aluminum oxide abrasive wheel
and a titanium alloy block. The sparks were thrown toward the surface of
the pool. Because of the dynamic nature of the test, determination of the
effects of variation in ignition source height was impractical. The
results of these tests are given in the table below.
Table 1
Average Minimum Fuel
Temperature for Ignition

JP-4 ~lBoF
Kerosene 1410 F
20
 200 -
Y-
o
Kerosene

150 - 8
c

-
0
.- ~
c
Dl

-
g
100 -
.
III

-..
:::»
g
III
A.
E
...
III
SO J P-4
III
:::»
Y-

E
:::»
E
c 0
~
III

..
g
Dl

III
>
c(
-SO -
Q 2 3 4 5
Height of Ignitor above Pool inches

FIGURE 12

The Effect of Hot Wire Ignitor Height on the Minimum Fuel

Temperature Required for Ignition of Unshrouded Fuel Pools

21
 The next sequence of experiments centered on determining the effect
of wind velocity on the minimum fuel temperature for ignition. -The
ignition sources were placed at the downstream edge of the pool, close
to the pool surface. Fuel temperature and air temperature were maintained
approximately equal and were increased at each wind velocity ;until ignition
occurred. The data obtained from these experiments are plotted on figure
13. The open flame ignition tests indicate that for any wind velocity
the kerosene ignition temperature is significantly higher than that for
JP-4. For both kerosene and JP~4, the minimum fuel ignition temperature
increases with higher wind velocities. The JP-4 spark ignition experi-
ments indicate that the effect of wind remains characteristically the same
regardless of the type of ignition source. The difference in the JP-4
ignition temperatures with the open flame and the electrical spark is not
attributed to convection (because of the wind velocities) but rather to
the difference in wind and fuel vapor flow patterns over the two geomet-
rically different ignition sources. The ignition of kerosene with
electrical sparks was attempted; however~ a facility limitation precluded
these tests. To achieve high fuel temperatures in a wind environment,
high air temperatures were also necessary. The wind temperature require-
ments needed in the kerosene tests were beyond the limit of the air
heating facility.

The final series of pool ignition tests were conducted with blends
of JP-4 and kerosene using an open flame as an ignition source. A ten
inch diameter pool was used~ with a propane flame held one inch above the
pool center. The average minimum temperatures for ignition together with
measured flash points (17) are plotted agaL"lst blend. composition in figure
14. These data show that the flash point of a fuel or a fuel blend is a
reliable indicator of the ignitability of fuel pools. Furthermore, it
is apparent that small additions of JP=4 to kerosene significantly change
the flammabi1ity~ while small additions of kerosene to JP-4 have a neg-
ligible effect on flammability.

In summary, the fuel pool ignition data indicate that ignition source
position, ignition source type~ and environmental conditions all affect
the minimum temperature required for the ignition of each fuel. The
difference noted between fuel types~ however, is primarily a function of
the flash point of the respective fuel.

Spray Mode

Various investisations have been conducted on the flammability char-

acteristics of sprays. Burgoyne (18) has shown that a mist with droplet
size below 10 microns behaves like a vapor; i t will have the same lean
flammability limit as a vapor and the same flame speeds for equal fue1-
air ratios. For mists with droplet size above 10 microns the lean
flammability limit decreases with increasing droplet size until a size
of 60 microns is approached. Anson (19) has demonstrated that the lean
flammability limit for a mist of 60 micron droplets is below that for 8
vapor, and has shown that for mists with droplet size above 60 microns
the lean flammability limit increases with increasing drop size.
22
 200 -
~
0

c
0
150
c
en - Open Flame

- a

..
GI
100
-.. ::t
a
GI
a.
E
... GI

50
GI
::t
~

JP-4 Spark
E
::t
E
.-c 0
~

GI
9
..en
D
411
6 JP-4 Open Flame
>
c(
-50
0 2 4 6 8

Wind Veloci ty mph

FIGURE 13

The Effect of Wind Velocity on the Minimum Fuel

Temperature Required to Ignite a Fuel Pool

23
 160

150 Flash Point (Pensky-Martens>

140
Average Minimum Fuel
Temperature at Ignition .•
130
·••••
•
•••
120 ••
•••
••
110 - •••
••
100 •••

0
Yo ·
•
•••
90 ••
•••
••

...a.
Q) 80 ••
~ b
••
70 •••
Q)
••
a. •••
E 60 •
Q) ••
t- •••
50 ••
•
•
••
40
D.'•
.•
•'
••

30 ••••••
.'
20 • •1
.'
••
~... 0
•••
10
......
..........
.....•••
0

10
o 20 40 60 80 100

Percentage of Kerosene in Blend

FIGURE 14
The Effect of Fuel Blend Composition on Flash Point
and the Average Minimum Temperature at Ignition
(Open Flame Ignition Source Placed 1 Inch Above Pool Center)

24
 Burgoyne, in his experiments with tetralin, shows that, for mists
with droplet size about 10 micrDns, flame speed increases with drop-
let size up to approximately 20 microns and then tends to level off
(there are-no data reported for droplet sizes above 38 microns). .
Eventually, as reported by Williams (20), flame speed should'decrease
with drop size when heterogeneous combustion becomes the controlling
mechanis~4

LiebIqBl1 (21) has reported data indicating that the energy required
to ignite a mist is about 25 times -greater than that required to ignite
the vapors of the ~ame material. The spark energy required to ignite
a mist varies with temperature and fuel type 0 Liebman's data for JP-4.
and kerosene are shown in figure 15, and indicate through extrapolation
that a constant ignition energy requirement of approximately 5 milli-
joliles exists at the respective flash point temperatures of the fuels.
Campbell (22) has investigated the ignition characteristics of JP-4,
kerosene and avgas sprays using friction sparks. Bearing loads of 20
to 1455 psi with sliding speeds of 5 to 40 miles per hour were obtained
with five different metals. The effect of bearing surface was studied
by using both concrete and asphalt runways.
The results indicate that aluminum will not produce friction sparks
that will ignite the fuel sprays under the test conditions mentioned
above. The titanium alloy (Ti 100A) ignited fuel sprays readily even
at low bearing pressures (20 psi) and slow sliding speeds (5 mph). How-
ever, the other three metals tested, magnesium alloy (FS 1), chrome-
molybdenum (SAE 4130) steel am AlSl 347 stainless steel, ignited the
fuel mist at slide speeds and bearing pressures less than those expected
in an actual aircraft crash.
Campbell states that kerosene fuel appears to be slightly less
susceptible to ignition under the test conditions than JP~4 or avgas.
The difference was in the number of times the spray would ignite; how-
ever, this difference was slight. The kerosene did not ignite as often
as JP-4 at low bearing pressures, though both fuels gave inconsistent
results at these conditions. It is expected that at fuel temperatures
higher than that tested (ambient) the kerosene and JP~4 would be equally
susceptible to ignition.
The type of surface (asphalt or concrete) had no effect on the
titanium alloy, but the asphalt surface required higher bearing pressures
and sliding speed with the chrome~molybdenum alloy. It was assumed that
the magnesium and stainless steel would give approximately the same
results as the chrome-~molybdenum alloy.
Since the more relevant data have been reported on the static ignition
characteristics of JP~4 and kerosene mists, tests in this program were
aimed at determining the relative flame speeds of these fuels and their
ignition c~aracteristics under dynamic conditions. The apparatus
25
 100
80
60
til
II
40
~
0

=
.-
E 20 Kerosene Spray

10
..••...
>0-

..
m
II
8.0
6.0
••
C
1&1

c 4.0
'.••••..
••••
.-
..
o
••••
c ••
•••• JP -4 Spray
m 2.0 ••
E
~

E
.-c 1.0
0.8
Gasoline Spray
0.6

0.4

0.2
-50 -25 o 25 50 75 100 125
Temperature of

FIGURE 15
The Effect of Fuel Spray Temperature on the
Minimum Ignition Energy Requirements (Reference 21)

26
constrUcted for, and employed in, the flame speed tests is shown
schematically in Appendix 1. Basically, a fuel mist, produced by a
pneumatic atomizing nozzle, is injected into the top of a four foot
long tube, and ignition is provided at the bottom of the tube by means.
of a heated nichrome wire assembly. Airflow through the nozzle is
regulated by varying the air pressure, and fuel flow is controlled by
changing the total head pressure of an external fuel reservoir.
Ionization probes, separated by a one foot interval, record the passage
of the flame. The upward flame speed is calculated by dividing the
probe separation distance by the time required for the flame to pass
from the first probe to the second (as recorded on an oscilloscope).
The ionization probes register flame passage by detection of a small
amount of ionization associated with the passing flame front. This
technique is especially suitable for flame propagation measurements
because of its extremely sho.rt response time. Based on information
contributed by the nozzle manufacturer (23) and other available infor-
mation (24), mean drop size was calculated to be between 20 and 40
microns. Mixture flow velocity in the tube was between 0.4 and 0.6
ft/sec. in all tests. It is realized that in any attempt to obtain
absolute propagation rates, apparatus effects must be considered;
nevertheless, these experiments can be used to determine the relative
flame speeds of fuels. In these tests a large percentage of the fuel
collects on the flame tube walls and drips down the walls to the bottom
of the tube. For an accurate determination of mist fuel-air ratio, the
wall drippings are measured and subtracted from the total fuel flow
rate. Since the air flow is known, the mist fuel~air ratio is then
determined. Figure 16 illustrates the flame propagation rates for a
range of fuel-air ratios, where the fuel temperature range is between
500 F and 600 F. Due to the random variables associated with flame
propagation in tubes there is a certain amount of scatter apparent in
the data - especially with the kerosene fuel. It can be noted that
there is no apparent relationship of propagation rate to mist fuel-air
ratio in these tests; nevertheless, the bulk of the observed rates lie
in a certain range. This range, therefore, can be used as a basis of
comparison between fuels, and for these tests the results indicate that
kerosene has a higher range of flame speeds than does JP-4.
Experiments to determine the effect of mist temperature on the flame
propagation through fuel mists were also conducted. Modification of the
flame tube apparatus provided for fuel mist temperature variation by
routing the fuel and air lines through a heat exchanger. The flame tube
is enclosed in a metal box, hence, the tube temperature was controlled
by passing air of the desired temperature through the box around the
flame tube. Since the insulated box was sealed and wall drippings could
not be measured, an accurate determination of mist fuel-air ratio could
not be determined. Although fuel-air ratio values of the non-ambient
test are not specifically kn~tests were run within the same range of
total fuel-air ratios as the ambient fuel temperature tests.

27
 30

JP-4
.;
CII
....:
III

- 20 -

..
CII
g 00
0 0
No
Propagation
ClI: 0

c 10 -
0 0000 0 0
0
g
In 0
g
D.
...
0
A.

0 I I I

0 0.1 0.2 0.3 0.4

Fuel-Air Ratio (By Weight)

30
c

Kerosene
u
CII
III

-
..... C
.: c
20

..
CII
g c
ClI:

c 10 -
..
0

g c
c
00
C
In
g C
D. 0
0
...
A.
0 I I I

0 0.1 0.2 0.3 0.4

Fuel- Air Ratio ( By Weight)

FIGURE 16

The Effect of Fuel-Air Ratio on the

Flame Propagation Rate Through Fuel Sprays

28
 0
The 'data obtained at fuel temperatures of OOF, 55 0 F and B5 F are
plotted on figure 17 in bar graph form, These data indicate that flame
speeds through rich JP-4 mists do not change appreciably with tempera-
ture; however, flame speeds through corresponding kerosene mists increase
significantly with temperature. This phenomenon can be explained in .
terms of the relative volatility of the two fuels. As a flame front
passes through a fuel mist, the fuel droplets are preheated by the flame
front. The individual droplets experience an increase in temperature
before the flame front reaction zone reaches the droplets. Because
JP-4 and kerosene have similar heats of reaction, their average flame
front temperature profiles will be the same. JP-4 has a distillation
curve approximately 200 0 F lower than kerosene however, and a significantly
greater portion of a JP-4 droplet will vaporize and diffuse during the .
preheating time as compared to a kerosene droplet. Thus, for an equiva-
lent mist fuel-air ratio 1 the effective vapor fuel-air ratio in the
preheat zone will be richer for JP-4 than for kerosene.
For the rich mist fuel-air ratios reported here, the effective-vapor
fuel-air ratio for the JP-4 mist would also be rich. Thus, increasing
the temperature does not increase the flame speed. For kerosene, in-
creasing the temperature will bring the effective vapor space fuel-air
ratio from a lean value towards stoichiometric fuel-air ratio. Thus, the
kerosene flame speed would be expected to increase with temperature, as
it does. Eventually, a mist temperature will be reached where preheating
would make the kerosene mist vapor space rich, but these tests indicate
tha5' for 3C-micron particles, this occurs at a temperature greater than
100 F. '
For lean mist fuel-air ratios, temperature increases have a different
effect on JP-4 flame speeds. A significant study of prevaporization on
mists of lower overall fuel-air ratios has recently been reported (25,26).
This investigation indicates that at lean overall fuel-air ratios, the
flame speed through the more volatile mist will be somewhat higher. Thus,
for lean overall fuel-air ratios, flame speeds would be greater through
JP-4 mists than through kerosene mists. In addition, at lean overall
fuel-air ratios, an increase in temperature will effect an increase in
flame speed for both JP-4 and kerosene mists. Furthermore, for both fuels,
flame speeds should be equivalent for mists of particle size below 10
microns. For such fine mists, droplets are completely vaporized in the
preheat zone and have been reported by others to behave as a vapor.

In summary, for rich mists kerosene has the greater flame speed; for
lean mists JP-4 has the greater flame speed. Overall, neither seems to
offer a substantial safety advantage considering flame propagation through
mists.
 .

40 Kerosene

u
Gl
'"
~
30 Kerosene
-
Gl
'vJ -0
0 a=: 20
c
0
..,- Kerosene JP-4 I I JP-4
0
01
0 JP-4
Q.

IlL
°. 10 J I 1 n
Gl
E I I I I I
0
u..
D
0
OaF 55°F 85°F
Sproy Temperature

FIGURE 17
The Effect of Fuel Temperature on the Flame Propagation Rate Through Fuel Sprays
 It was noted in the ambient mist temperature tests that a rich
flammability limit existed for JP-4 at a fuel-air ratio of .024. With
kerosene mists J no rich limit was observed. This can be explained when
one considers the effective vapor fuel-air ratio mentioned in the dis-
cussion above. Furthermore, it was observed visually that the flame
front propagating through the kerosene-air mixtures was substantially
thicker than that propagating through JP-4 air mixtures. This indicates
that the kerosene droplets are not pre-vaporized as readily by the
approaching flame front, and in fact are actually post-vaporizing and
burning behind the flame front. Without this pre-vaporization the
effective vapor-air mixture is lean and explains why a rich limit was not
achieved.
An attempt was made to find a lower temperature limit for the
ignition of fuel mists. At the lowest temperat ure attained (-40 0 F), the
hot wire ignited both JP-4 and kerosene. Zabetakis (27) has reported
tests in which sprays of both kerosene fuel and a "wide cut" fuel were
ignited at -100 oF with a high energy spark. It can be concluded that
mists of JP~4 and kerosene can be ignited in any temperature environment
encountered in an actual crash.

In the final series of tests the hot surface ignition characteristics

of kerosene and JP-4 sprays were investigated. The test apparatus used
was the same'1 as that used in the stream tests, and a limited quantity of
fuel (5.9 cm )was sprayed onto the hot plate. The only difference be-
tween these two series of tests was that a low flow nozzle tip was used
to atomize the fuel for the spray tests whereas an orifice was used to
produce the stream. The data obtained are plotted on figure 18 along
with the stream data and indicate, as in the tests with fuel streams,
that for higher wind velocities there is a corresponding increase in the
surface temper ature required for ignition. Surface temperature ignition
requirements, for both fuels, are higher for sprays than for streams at
all wind velocities. This is apparently due to the fact that the wind
can carry the spray away from the heated region above the plate. The
sprays then have a lower residence time than the streams and therefore
the time allowed for ignition is decreased. This decrease in allowable
ignition delay time increases the thermal energy requirement for ignition,
and therefore a higher surface ignition temperature is needed. These
effects are explained in more detail in the stream discussion section
of this report.
Although there is a difference in the minimum surface ignition
temperature between JP-4 and kerosene, the magnitude of this difference
is not considered significant.

31
 1450 - _ - - - - - - - - - - - - - -.... --------------""'!.

••
1400 J p- 4 Spray ~
•••••••• P - 4 Stream

.... ••
o ••••
••••• ••••••
•••• ••
..
GI ...•• .•...••••
~ 1350
..a
GI
•••••••••• ••••••••
Q, .+ .+
••• ••• Kerosene Stream
E
...
CIl
Kerosene
Spray ••
•••••••
•••
••••••

c
o 1300 - ••• •••••
••• .+
c ••••
••• .+
.+•
m
CIl ••••••••••••••

-..
a
~
u

\It
1250 -
••
••

E
~

E
c
:E 1200-

1150
o 2 3 4 5 6 7
Wind Velocity ft.l sec.

FIGURE 18

The Effect of Wind Velocity on the Minimum

Surface Ignition Temperature for Fuel Sprays

]2
 CONCLUSlOOS

This investigation has provided information on the crash fire hazards

of aviation kerosene and JP-4 which substantiates the following
conclusions:
(1) The possibility of a flame climbing a column of dripping fuel
(JP-4 or kerosene) is extremely remote.
( 2) The minimum fuel temperature f or flame climbing up a stream
of fuel is significantly lower for JP-4 than for kerosene.
(3) The spatial ignition envelope, for fuel streams, ceases to
exist when the fuel temperature falls below the respective flash
points.
(4) The fuel pool ignition characteristics are affected by the
ignition source placement, the ignition source type, and environmental
conditions; however, the difference noted between fuel types is pri-
marily a function of flash point.
(5) There is no significant difference in the wind velocities
necessary to blowout pools of burning JP-4 and kerosene.
(6) Neither JP-4 nor kerosene offers a substantial safety advantage
when considering the flame propagation rate through mists and mist
ignitability.
(7) The surface temperature required for the ignition of both mists
and streams was not significantly different for JP-4 and kerosene.

33
 APPENDIX A
The test apparatus and apparatus configurations employed in this
test program are illustrated in figures 1.1 through 1.10. All drop, .
pool, and stream experiments as well as the fuel spray hot surface
ignition tests were conducted in the wind tunnel shown in figures 1.1
and 1.2. The wind tunnel configurations used for the stream, pool,
and hot surface ignition tests are schematically and photographically
illustrated in figures 1.3 through 1.8. A flame tube, designed to
measure the flame propagation rates through fuel sprays is shown in
figures 1_9 and 1.10. .

1-1
 @- Thermocouple
Thermocouple Station
TC

Conditioned
Air

•
I--'
I
1'0 LT::ct~IO:

Duct Length - 15 feet Air Velocity Range - 0 to 20 mph

0 0
Duct Diameter - 43 inches Air Temperature Range - -40 F to 130 F

Duct Pressure - Ambient

FIGURE 1-1
Schematic of Wind 'I'urmel Test Section
 gj
+>
'Po"
~

(\/
~
'"
..!J
+>
~
III
Q)
Eo-<
r:1 rl
'"'
J
'0
.~
:;:

1-3
 Fuel In

-----------~-l------------
Needle Tip _ _ .-J/ I
I
Stream
:~Fuel
Conditioned Air
t
I

Fuel Out

Fuel In

.}-------
Needle Tip - - . - / :
I
I
I
r Electrical Spark Ignitor

m~1
... ~ ( ignitor position varied)
"
Fuel St...

'\/
I
•
Fuel Out

ELECTRICAL SPARK IGNITION CONFIGURATION

FIGURE 1-'3

Schematic of the Test Section Configuration

for Stream Ignition and Flame Climbing Tests

1-4
 FIGURE 1-4
Burning Stream after Ignition and During Flame Climbing

1-5
 Fuel Source

---------]~---­
Wind Ir Fuel ln i8ctor

AI, Flow Guide Pla7 k:/ ot Plate

____~ ~J=H=ea=~:=~=(~ _

FIGURE 1-5

Schematic of the Test Section Configuration

for the Hot Surface Ignition Tests

1-6
 "
<l
+'
'"
H

~'"
-.0 §
~ OM
+'

~
'M
SJ
~
H
rx. Q)
()

'"
'H
8
Ul
+'
0

""

1-7
 Fuel Pool

SHROUDED CONFIGURATION

Ignition Source

UNSHROUDED CONFIGURATION

FIGURE 1-7

Schematic of the Test Section Configuration for Pool Tests

1-8
1-9
 Fuel Air

Reservoir Supply

Fuel Flow Meter . .

M
Fuel Nozzle
Pressure
Regulator

Temperature Control _ ..._ _-+~

Air - In

Flame Tube
/1\
' \ - Ionization
Tube 'nd..u•• ~ I Probes
-+--+---0

Thermocouple CD --+--
Thermocouple

Temperature Control
Air - Out

Ignitor

FIGlJRE 1-9

Schematic of the Apparatus Employed for the Flame Speed Tests

1-10
 FIGURE 1-10

Flame Speed Test Apparatus

1-11
 APPENDIX B
The following table lists the typical properties of the test
fuels employed~

FUEL JP-4 Kerosene

Gravity, Specific, 60/60 of 0.7567 0.8309
Gravity, °API, 60/60 of 55.5 38.8
Reid Vapor Pressure, lb/in 2 2.60 0.13
Distillation, I.B.P. of 160 350
5% over of 186 380
10% over of 194 388
20% over of 208 400
30% over of 220 412
40% over OF 234 422
50% over OF 250 432
60% over OF 260 444-
70% over OF 280 456
80% over OF 310 468
90% over OF 370 484
95% over OF 438 500
End Point OF 454 504
Recovery % voL 98.5 98.0
Residue % vol. Ll 1.2
Loss % vol. 0.4 0.8
Sulfur, % wt. 0.15 0.36
F.I.A. Saturates, % vol. 75.00 75.34
Olefins, % vol. 2.63 5.47
Aromatics, %vol. 22.37 19.18
Aniline Point, °c 49.0 6106
Aniline - Gravity Constant 6,671 5,544
Heat of Combustion, BTU/lb 18,668 18,451
Freeze Point, OF -52 -60
Flash Point, OF -6 1.42
Viscosity, cks., 100°F 0.64 1.75
-30°F 1.74 12.88

2-1
 APPENDIX C
REFERENCES
L "Tangents," SAE Journal1§. (5), 1970,
2, Pinkel, 10 10, Weiss, S., Preston, Go M" Pesman, G, J.,
"Origin and Prevention of Crash Fires in Turbojet Aircraft, II
NACA TN '397" May 1957,
, , Pinkel, 1. 1., Preston, G, M, Pesman, G. J., "Mechanism of
Start and Development of Aircraft Crash Fires," NACA REP. 1133, 1953.
4, Scull, W, E" "Relation between Inflammables and Ignition Sources
in Aircraft Environments," NACA TN 2227, December 1950.

5 Levin J S, M., "Surviving the Crash," Space/Aeronautics !fl (5),

p. 88-99, 1968,
6. Spalding, D. B., "The Combustion of Liquid Fuels," Fourth
International Symposium on Combustion, p, 847-864, Baltimore:
The Williams and Wilkins Company, 1953.
7. Isoda, H., Kamagai, S., "New Aspects of Droplet Combustion," Seventh
International Symposium on Combustion, p, 52'-5,1, London: Butterworths
Scientific Publications, 1959.
8, Goodall, D, G., Ingle, R., "The Ignition of Inflammable Fluids by
Hot Surfaces," Aircraft Engineering 38 (4), p. 20, 24, 29, 35, 1966.
9, MacDonald, J. A., "Ignition of Aviation Kerosene by Hot Pipes,"
Royal Aircraft Establishment TR 67162} 1967. AD 824289
10. MacDonald, J. A., Royal Aircraft Establishment, Private Communication,
May 1970.
lL Glassman, 10, Hansel, J. G" "Some Thoughts and Experiments on
Liquid Fuel Spreading, Steady Burning J and Ignitability in Quiescent
Atmospheres," Princeton University Aerospace and Mechanical Laboratory
Report 841, July 1968.
12, British Ministry of Aviation, "Report of the Working Party on Aviation
Kerosene and Wide~Cut Gasoline," C.A.P. 177, 1962.
1" Blinov, V, I., Khudiahov, G. N" "Certain Laws Governing Diffusive
Burning of Liquids," Acad. Nauk, SSSR Doklady, 113, p, 1094, 1957.
14, Kuchta, J. M" Furno, A. L., Martindill, G. H., Imhof, A. C., "Crash
Fire Hazard Rating System for Controlled Flammability Fuels," FAA,
NAFEC, Report No. NA-69-l7, March 1969. AD 684089

3~1
15· Walker, J. R., "The Effect of Wind on Uncontrolled Buoyant
Diffusion Flames from Burning Liquids," Ph, D. Thesis,
University of Oklahoma, Norman, Oklahoma, 1965.
16, Hirst, R., Sutton, D., "The Effect of Reduced Pressure and
Airflow on Liquid Surface Diffusion Flames," Combustion and
Flame ,i, p. 319-330, 1961.
17, "Standard Method of Test for Flash Point by Pensky-Martens
Closed Tester, II ASTM Standard D93-66, Part 17, p. 24-36,
American Society for Testing Materials, Phila., Pa., 1969.
18. Burgoyne, J, H., Cohen, L" "The Effect of Drop Size on Flame
Propagation in Liquid Aerosols," Proceedings of the Royal Society
of London, Series A, Vol. 225, p, 375=392, 1954.
19, Anson, D., "Influence of the Quality of Atomization on the
Stability of Combustion of Liquid Fuel Sprays," Fuel '32, p. 39-51,
January 1953,
20, Williams, F. A., "Monodisperse Spray Deflagration," Progress in
Astronautics and Aeronautics, Vol, 2, p, 229-264, New York:
Academic Press, 1960,
21, Liebman, I" Spolan, I" Kuchta, J. M., Zabetakis, M. G.,
"Ignition of Tank Atmospheres during Fuel Loading," API Division
of Refining Meeting, Paper No. 36-65, API Proc, 45 (3) p. 218-230,
1965.
22, Campbell, J, A" "Appraisal of the Hazards of Friction-Spark
Ignition on Aircraft Crash Fires," NACA TN 4024, May 1957.
23, Spraying Systems Company, Bellwood, Illinois , Private Communication,
April 1970.
24, Fraser, R. P., "Liquid Fuel Atomization," Sixth International
Symposium on Combustion, p. 687~701, New York: Reinhold Publishing
Corporation, 1957
25, Wood,B, J" Rosser, W, A., Wise, H" "The Ignition and Combustion
of Drops in Sprays of No.2 Heating Oil," API Research Conference
on Distillate Fuel Combustion, Conference Paper CP 63-67, 1963.
26, Wood, B. J" Rosser, W, A" Wise, H., "The Effect of Fuel Droplet
Prevaporization on Flame Propagation Velocities," Combustion and
Flame 11, p. 663, 1969.
27. Zabetakis, M. G., Jones, G. W., Scott, G. S" Furno, A. L.,
"Research on the Flammability Characteristics of Aircraft Fuels,"
WADC Technical Report 52-35, Supplement 4, p, 21, January 1956,
.3-2