Publisher’s version / Version de l'éditeur:
Vous avez des questions? Nous pouvons vous aider. Pour communiquer directement avec un auteur, consultez la
première page de la revue dans laquelle son article a été publié afin de trouver ses coordonnées. Si vous n’arrivez pas à les repérer, communiquez avec nous à PublicationsArchive-ArchivesPublications@nrc-cnrc.gc.ca.
Questions? Contact the NRC Publications Archive team at
PublicationsArchive-ArchivesPublications@nrc-cnrc.gc.ca. If you wish to email the authors directly, please see the first page of the publication for their contact information.
https://publications-cnrc.canada.ca/fra/droits
L’accès à ce site Web et l’utilisation de son contenu sont assujettis aux conditions présentées dans le site LISEZ CES CONDITIONS ATTENTIVEMENT AVANT D’UTILISER CE SITE WEB.
Report (National Research Council of Canada. Division of Building Research), 1954-12-01
READ THESE TERMS AND CONDITIONS CAREFULLY BEFORE USING THIS WEBSITE. https://nrc-publications.canada.ca/eng/copyright
NRC Publications Archive Record / Notice des Archives des publications du CNRC :
https://nrc-publications.canada.ca/eng/view/object/?id=f84b6e94-e1cf-47be-869e-429240f1b9f5 https://publications-cnrc.canada.ca/fra/voir/objet/?id=f84b6e94-e1cf-47be-869e-429240f1b9f5
NRC Publications Archive
Archives des publications du CNRC
For the publisher’s version, please access the DOI link below./ Pour consulter la version de l’éditeur, utilisez le lien DOI ci-dessous.
https://doi.org/10.4224/20337849
Access and use of this website and the material on it are subject to the Terms and Conditions set forth at
Report on a Model Test to Study a Fire at Scarborough, Ontario
Burnett, C. G.
NATIONAL fュセZ[イZjarciMi C;OUNCIL C1UTADA
REPORT ON A MODEL TEST
TO STUDY A FIRE AT SCAHDOHOUGH, ONTARIO
by
c.
G. Burnett
ANAl YZED
Not for Publication
Report No.
38
of the
Division of Building Research
Ottawa
December
1954
I
I
IHTHOLUCTIOlJ
During actual firefighting, members of fire departments seldom have the opportunity of
s e ei ng the rapid spread of a fire wheri it is
in its first stages. Information about flame
spread can be of 3reat assistance to the
fire services and buildinc officials in their vJOrk of pr-o t.e c t.Lng life and property.
This report outlines the model test made to
study a fire in whi ch hIO childron 10 st their
lives. As such tests reveal much helpful
information, it is hoped that it may aid in averting trusedies in the future.
'r'. -. -n--, ",[11 '1 "';\I
ャャNゥNZjエセiセlNl l:J.'J fj"lT:"-'_--1' ..ciセMゥ1 ... nn.l.v
, f i
iLL
by C. G. Burnett
On Wednesday, March
3, 1954,
atE
a.m., a fire occurredin a one-room IO-foot square dwelling at Highland Creek, north
of .lo , 2 Ei;hway in the 'I'own ahLp of R」。イエッャセッオR[ィ rie ar- 'I'or-orrt.o ,
The dwe Lling was occupied by a lirs. 'diLdns with her three
chl Ldr-en aGed セ⦅L 2, and 1 year, and a Ers. Hampson ,:Jith one child
3
years old. The fire was responsible for the death by suffocationof the t wo children aged
3
and Qセ year s and caused severe burnsto three other occupants of the dwelling.
On the day following the fire, an investigation was
conch: cted at the scene by C. G. Burnett and P. Huot from the Fire
Research セ・」エゥッョ of the Division of Buildins Research in order
to ascertain the cause for the rapid spread of fire within the 「オZゥlャゥZZQ{ェセ
'This dHslliw;ms 3. one-storey buLl.d Lnr; of' frame
construction, supported 0;" co n-: "ate bLoe ks , Combu st LbLe board
was used for both interior ana exterior sheeting. Insulatlon
was placed over the ceiling between the roof joists, but
there was no insulation in the walls. The combustible board
ヲッイセゥョァ the interior surface of the walls and ceiling was
painted.
2.2 Services
Heating and cooking facilities were provided by a stove, located at the north end of the building, built to use wood or coal but converted into an oil-burner type with
a blower placed at the rear of the stove. A two-gallon
galvanized tank was attached to the north wall at the west side of the stove with a copper pipe leading from the tank
to the carburetor and into the combu st Lon chamber of the
stove. 'I'h e building was wired for electricity. There was
no water service to the dwelling.
2.3
Living AccommodationSleeping accowJodation was provided by a double bed, t1rJO cribs and a folding bed; the latter was removed in the
daytime to provide more space in the room. The kitchen
section was on the east side of the building where a
refri3erator and table were located. A wooden counter with
shelves to the floor, vrhich were used to hold the sink and
store various other articles, had been constrllcted on the
south wall near the entrance. A wire clothesline ran from
2
-2.11 Detection
The fire was discovered at
S
as m, by HI's. Wilkins whostated that she was facine south in the room, dressing, and
on turning around noticed fire in the vicinity of the oil
tank and spreading rapidly up the セBAゥョ、ッh curtains and vraLl,
to the ceilinG.
2.5
Action of OccupantsOn di scovering the fire, r':rs. \'Jilkins pLc ke d up the
two younger children who were in their cribs, telling the older children playing in the room to follow her outside. The rapid spread of the fire across the room to the entrance of the dwelling prevented Mrs. Wilkins from re-entering the
bul Ld Lng , (lIrs. Hampson had left for work at
7.30
。NセョNLleavin::; llrs , l-Jilkins alone with t he children).
2.6
ExtLn-rud\r,( shmerrtThe fire was vext.Ingui sbed Lmncdl at ely after the ar-r-I"val
of the fire department. The bodies of the two older children
were found suffocated underneath the 、oセjャ・ bedo
2.7
Oripin and Cause of FireExanlination of the buiLdLrig by members of the lJivision
revealed that there was a cardboard box containing a can,
placed near the base of the oil tank, trhLc h contained some oil
that had dripped from a leaky oil line. It is difficult to
ascertain the actual cause of the fire but the following
ass1L'nption is made: a ciGarette or match was thrown on to
the oil-soaked floor or into the cardboard box. The fire
could have smoullered and bur st into flame whi Le
r-rr
s , 1,'ilkinswas turned away from the stove. Doth HI'S. セjゥャォゥョウ and HI'S.
Hampson admitted lighting ci arettes in the dwelling.
The fire spread very rapidy, igniting the curtains
and the combustible board. Articles of clothing and other
combustible material in the room aided the fire in its rapid
spread towards the only exit from the building. The fire
involved the entire room and contents before being extinGuished.
3.
MODEL BUBNIn order to study with Hhat rapidity this fire spread in this type of construction, it was decided to construct a model
and carry out a model burn. It was not possible to scale all
dimensions accurately but approximately
4
scale was used. A3
-3.1
Description of Model and Contents(a) nodel
The model and its cant ent s were constrncted in the
shop of the Fire Research Section, Division of Building
Research, National Research Council (Figs. 2 and
3).
Thedimensions of the model were
4
by4
by 2 feet with wallsand ceilina constructed of 1-inch wood fibreboard and
fLoo r Lng
of
i-inch Do'.:glas fir p Lywoo d ,'I'wo TTindows were nlaced in the bu l Ld l ng , one h by
12
inches was located in the west wall and the other,6
inches square in the north wall. Both windows were fitted
with s Lng I.e diamond slass fixed with scotch tape to the
exterior of the wall. Entl.ance to the building was from
the east side with a door made of !-inch combustible board hung to open Lnwar-d ,
(b) Furniture
Four metal beds were constructed and suitably located.
Samples t aken from the orir;inal mattress were placed on the
beds and covered with cloth to simulate sheets. A refrigerator
and table made with t-inch plywood were placed along the east
wall. A counter and shelf constructed of ':i-inch plywood were
located at the entrance to the dwelling along the south wall. A metal model of a stove was placed at the north end of the
building. In addition a 3-inch liameter セィ。ャャッキ cup filled
vJith 1 ounce of oil was placed on a newspaper on the floor
beside the stove to simulate c ond Ltions exd st Lng at the time
of the fire.
Cloth curtains were hung from both windows and a wire clothesline was attached to the interior walls leading from
the stove. Pieces of cloth were attached to the line in order
to determine if possible whether the spread of fire was influenced by the articles hung on the line as well as the window curtains.
3.2
Weather conditions on March31,
1954.
Wind Air Relative
Time Velocity Tempera- Humidity
m.p.h. ture (OF)
(;;
)3.15
p.m. vJNW2-6
43
43
It
-The relative humidity ana air temperature were recorded
on a Bendix-Priez hygrothermograph. The wind velocity
was obtained by using a Biram type anemometer.
3.3.
Procedure(a) Ignition
The oil in the porcelain cup was ignited by using a steel rod with a lighted piece of cotton attached to the end and extending it through the door entrance.
(b) Gas sampling
During the course of the burn, samples of the
combustion gases were taken at specific times and locations in order to obtain some picture of the rapidity at which the atmosphere deteriorated inside the model.
The apparatus used for this purpose is illustrated
in Fig.
I!.
It consisted of eight pre-evacuated samplingtubes connected to a distributor line which, in turn, was fed from three different locations in the model by means
of rubber tubing and steel piping. A system of valves
installed between the distributor and the feed lines
permitted the ウ。セーャゥョァ of gases from one location at a time.
The other end of the distributor was connected to a suction system, provided by a syphon, which allowed the continuous removal of the gases in the line at a rate, measured by a
flowmeter, of
15.5
litres per minute. The time lag betweenthe withdrawal of the gases from the model and their collect-ion in the s arnpLd ng tubes was calculated from the flow rate and was found to be approximately 10 seconds. Each sample could, therefore, be taken as representing the mean con-centration of the gases over a period of approximately ten
seconds prior to the actual time of sampling. It is thus
assumed that, even though the stopcock to each evacuated
tube was left open for a period of 10 seconds, the gases would rush in as soon as the stopcock was opened.
The actual setup of the apparatus during the burn is
shown in Fig.
5.
The exact locations at which samples weretaken are noted on the general drawing (Fig.
6),
and wereas follows:
Location A: Under the double bed at a point one inch
above the floor,
8
inches from the rear wall and 10inches from the left-hand side セ。ャャ[
Location B: Above the double bed, at a point directly
above Location A at a ィ・ゥセィエ of
15
inches from the floor;Location
c:
Near the door, at a point15
inches abovethe floor,
8
inehe s inside the door and 10 inches f-'o!llGas s ampLes 1rIers dr-awn l'ss;ulal'ly f'r-orn location A at ons-md.nut.e inter-vaLs , 1'hi::; procedure \-JElS adopted in the
expectation that G. minimum of five samp Les could be obtained
beI'ore the I'Lashe o ver- occur-r-ed, 7his as sumpt Lon Has based
on the results of tests conluc.ted in Great Britain
(1)
whichshowed that, with rooms lined with wood fibreboard, the
flasl'J-over h ad t al.er, pLac e approximately
S
minutes afterienition. However , in the pr-cscrrt case, t.h l s phenomenon
occurred at about 21 r,lircutes bec acs e ignition lIIaS started
wi til oil in order t o siy:mlate as closely as possible the
conditioYlswhich existed at the actual fire. Only three
s ampLes セGj・イ・ obtained f'r-orn location A at orie-emlnut e intervals,
starting f'r-orc the time of iGnition. THO other samples were
collected, one each from locations Band C in order to determine the approxi"':.ate distribution of the ccrr:.bustion
pr-odu ct s in the model. In each case, the sample Li ne Has first
connected to the suction system for a minimLlin period of 20
seconds to clear the 0istributor of the gases previollsly
wi t.hdr-awn, At the non.Lna.l time of s amp Ld ng , the stopcock
to one of the sampling tubes was opened for a period of 10
seconds arid then closed. The s arnpLes thus obtained Here
subsequently analysed.
( c ) 'I'emnez-a* t ur-e ゥᄋセ・。ウョイXュN・ョエ - .__. _ - - - -s
of the carbon type.
Terrperatures were measured and recorded using
22-ganee chr-omoLe-aLurne L t.her-mccoupLes connected to a multiple
point (16-point) recorder (See Fi;.
5).
Thermocouples werelocated at
IS
dLf'fer-errt points as 811.01,vn in FiG.6.
Thethermocouple s installed near the fLoo r and near the ceiling
'Here enclosed in fiberGlas s l.e evLng and those at rnid·uheight
were enclosed in 2-hole porcelain insulators. The recorder
was started as soon as the oil in the container near the stove ignited.
4.
RESULTSセ .• l Analysis of the combustion ァ。ヲ_セセ
The five smnples of gases collected durinG the conduct
burn Here analysed for carbon dioxide, oxygen, and
monoxide with a portable gas analyser of the Orsat The results of the analysis are given in Table I.
- 6
-TABLE I
Nominal Time
°2(%)
CO
2of Sampling (min.)
Location
(%)CO
0&)
1.05 A 20.8 0.0 0.0 2.00
A
20.8 0.0 0.0 2.30 B 4.9 12.7 5.53.00
A
19.6 0.8 0.6 3030C
20.6 0.0 0.04.2
Temperature in model during the burn
The temperatures obtained during the model burn up
to the point of exting1"lsbmem. are
ウィッセュin Table II.
TABLE II
Thermo-couple
Time
Tern)"
Time
Tem) "
Time
Tern).
Time
Temp.
No.
(
sece
,
J(OF
(sees)
(OF
(sees)
(OF
(sees)
(OF)
-'. セ
---I 32
Nil
96Nil
160Nil
2
t
6 60 100 228QVセ
935セ
セッ 85QPセ
205 16 955エセ
82 10 250 172 9205 81 112 230 176 1065 6 52
Nil
116Nil
180Nil
セ
56 275 120 575 18k 1080 60 260QRセ
595 18 13409
セセ
355 12 1110 192 1190 10セ
554
30 132 720 196 1150 11 55 72 90QセV
915 200 1070 12 12 55 76r
5
1+0 845RPセ
10f5ゥセ
16 70 80 '20Qセセ
1130 20 13,0 20 95セセB
20 14 1100 212 1310 15 24 98 555 152 1065 216 1280NOTE:
a)Point No.
16was not used.
b)
No readings obtained for points
1and
6.5.
DISCUSSION
5.1
Combustion Gases
The results of the gas analysis tend to show that the
composition of the atmosphere inside the model varied a great
deal not only from one level to another, but also from one
7
-location to another at the same level. Visual observations
as well as temperature measurements clearly indicate that the
most severe burning took place at ceiling level. This and
the effect of convection on the combustion gases Hould explain the high concentration of CO and C02 detected at
location
B
o On the other hand, at pointC,
situated at thesame level as point B, no evidence of' combustion can be
obtained from the gas analysis althouGh it was quite evident
from visual observation that some severe 「オイョゥョセ occurred
at that location. The incompatibility of these findincs
could be explained by the two following facts:
(i) Point C was situated right above the door opening
so that a continuous draft of fresh air could have
affocted the compo si tion of the atmo sphere expected
at that point;
(ii)
The sample of gas collected from pointC
was drawnHhile the extinguishment of the fire, throuGh the
door openinG' was actually in progress.
The analysis of the samples collected at floor level under the bed indicates that very little conillustion occurred at that point, which is further corroborated by the temperature readings and also by the fact that, after the fire had been
put out at about the SセMュゥョオエ・ marks the floor at point
A
showed hardly any visible sign of burning. A similar
observat-ion had also been made at that locatobservat-ion at the actual scene
of the fire. The first two gas samples proved to be of the
same composition as that of the atmosphere, that is, about
21 per cent l'xygen by volume. The sample taken at the エィイ・・セ
minute mark (after the f Laah-vover ) shows, however, a low
concentration of combustion products in the same area, VJhich could result either from some actual but very slow combustion
or from the diffusion of CO and coセ from other
parts
of theof the model.
From the 。「ッカ・セ it can thus be concluded that:
(i)
(ii)
Up to the three-minute mark, and possibly for some
time after, had the fire been left burning the oxygen concentration at location A (under the bed) Has quite
sufficient to support life
(2);
At the エィイ・・セュゥョオエ・ mark, the concentration in carbon
monoxide was in the range at which "most animals would
die in a very short time"
(3).
It is believed thatthis concentration increased as the burning progressed.
5.2
Room t・ュー・イ。エオイ・セIn summarizing the temperature conditions during the model burn, reference is made to Fig. 16 which shoHs the
8
-heights above the model floor throughout the test. Some
of these relationships are set out in Table III. TABLE III
Time Average Temperature in Nodel in
VF
(sees) Ib above floor 12 above floor H3 above floor
60 100 365
90 210 AセYP 585
120 lセャo 705 830
150 740 925 1030
It can be seen from Table III that conditions inside the model would have been quite intolerable shortly after
one minute from ignition due to the excessive heat. In
addition, the room was fully involved (flash-over point)
at around Rセ minutes. This rapid spread could be accounted
for by the use of fuel oil and light combustible material in
the test. During other model burns carried out by the
Division of Building Research with the same type of wall
lining, the model room has become fully involved in Tセ to
5 minutes. In these tests no additional fuel was added in
the form of furniture, textiles, or fuel oil.
1rJhile no definite conclusions may be reached based on this test, i t does serve to show that the mother had very little time to carry out any rescue operations and points out the foolishness of using a leaky heating device utilizing a liquid fuel in a room lined with combustible
material. This factor, coupled with crowded living conditions,
was undoubtedly responsible, in part for the tragedy. HEFERENCES
FoE.T. kゥョセュ。ョL EoH. Coleman, and D.J. Rasbash.
Chem., 3, 463-68 (1953).
J. Applo
2. Y. Henderson and H.W. Haggard. Noxious gases and the
principles of respiration influencing their action. The
Chemical Catalog Co., N.Y., p.37 (1927).
N.F.P.A. Handbook of Fire Protection,llth Edition, Boston, 1954, p. 346.
II
48
DOUBLE BED CRIB CRIB
D D
D
FOLDING COTD
CUPBOARD OIL P'CiAND SINK OIL
stovセ
7
I
.:
ICE KITCHENbセ
(
TABLEFLOOR
PLAN
OF
MODEL
FIGURE
Fig. 2
Interior view of model
Fig.
3
Interior view of model
RUBBER TUBING
Y4"
1-0-STEEL PIPING '1 3/1'6 1-0-SUCTION LINE 01 STR IBUTOR I •• Y'Z ' I- O· SAMPLING BOTTLES VALVESFIGURE
4
GAS
SAMPLING
APPARATUS
Fig.
5
Test apparatus and model
AB
13 2 8 0 o 1 3 12 0セ
51015 C 4914 0 0 1611PLAN
II 14 0 II 120 120CB
015 13 14 0 01& 13.C
-B
60 010 9 0 ' 0 010 10 1 8 9 8 10 0& 4 0 10 0& 20 2 A 3 4 3.A
E.LE.,vATIONE..LE..V.
A
Fig.
6
Location of thermocouples and gas sampling tubes
Fig.
7
Thirty-eight seconds after ,i gn i t i on .
(The oil in the evaporating cup on the floor at '
the north wall has ignited and the fire h8s
exten-ded to the window curtains)
Fig.
8
One minute,
16
seconds after ignition.
Hセィ・
fire is advancing rapidly up the north wall to the
ceiling.
Heat broke the glass held in place with
scotch tape.
The tape was removed when the glass cracked)
Fig.
9
One minute, 31 seoonds after ignition.
(The fire has advanoed up the north wall, extending
aoross the room to the door entranoe by way of the
oeiling and floor)
Fig. 10
One minute,
50
seoonds after ignition.
(Heat and gases are being generated within the
dwell-ing.
Flames are attaoking the outside of the bUilding
through the window)
Fig. 11
Two minutes, 15 seconds after isnition.
(The fire has extended involving the entire room)
Fig. 12
Two minutes,
50
seconds after ignition.
(Flames now engulfing the interior of the building)
0 c
.
E CD CD tit•
セ ::J 0 0,..
Co) 0 E..
N•
t-セ t- O No
o
No
o
oo
o
II)o
i
o
o
,..
s
No
.:I0 SャエョャyセS、nSQThermocouple. '2" from floor T8 200 1 1 ...4 I I I I I I I I 1 1000 I I T f0 1200
I I
T7I
I
I
I
=t 1 74=
pt-1
1400i i i i i i , i i LL. 0 8001 1 T9...
ar: :J.-
600«
ar:...
Q. セ...
.-
400 0 20 40 60 80 100 120 140 160 180 200Imin 2 min 3 min
TIME IN SECONDS
FI GURE
14
2001 I / " セ ...セ I I I I I I I I
FIGURE
15
o
20 40 60 1min. 80 100 120 140 2 min. TIME IN SECONDS 160 180 3 min. 200 220 De.P, RE.pog,T 38180 200 3min 160 140 100 120 2mln TI ME IN SECONDS 80 60 Imln 40 20 I
I
I
AVERAGE TEMP. 18" ABOVE FLOOR
jFLASHOVER 0