Ulf Wickstrom^

Application of the Standard Fire Curve

for Expressing Natural Fires for
Design Purposes

REFERENCE: Wickstrom, U., "Application of the StandanI Fire Curve for Expressing
Natural Fires for Design Purposes," Fire Safety: Science and Engineering, ASTM STP
882, T. Z. Harmathy, Ed., American Society for Testing and Materials, Philadelphia,
1985, pp. 145-159.

ABSTRACT: A method suitable for design purposes is presented which allows postflash-
over compartment fires to be expressed in one single curve; time is then modified or scaled
to take into consideration ventilation conditions and wall properties. Ventilation-con-
trolled combustion, uniform temperature, and wall losses proportional to the thermal iner-
tia of the wall material are assumed. Although the proposed method allows the designer to
consider important parameters, it is still very approximative and must be used with care.
Several assumptions are made, and many parameters are difficult to assign values in the
general case. The method therefore should be looked upon as an improvement of the cur-
rent technique of carrying out fire engineering design by simply employing a standard
curve like ISO 834.
An analysis of the postflashover compartment fire, basically the same as has been pre-
sented elsewhere, is summarized. The parameters of ventilation and wall properties are
grouped together in such a way that their influence on the temperature development may
be expressed as a time factor. It is further shown that for certain parameter values the fire
temperature response in the heating phase nearly coincides with the ISO 834 standard
curve. Therefore, as this curve is well-established in fire engineering, it is suggested that it
be employed with the here suggested method; the time scale, however, is modified with a
compartment time factor depending on the properties of the fire compartment.
The assumed fire temperatures may be used to estimate the temperature in fire-exposed
structures. It is demonstrated how very compact diagrams may be used for concrete struc-
tures, which allows the designer to take into account various fires and thermal material
properties of the structure. The diagrams were obtained from finite-element computer cal-
culations in which various fires and material properties were simulated. By then plotting
the temperature response versus suitable parameter groups, it was possible to present the
results in a very compact way. Corresponding diagrams have earlier been presented for in-
sulated steel structures [/].

KEY WORDS: compartment fire, postflashover, temperature calculation, concrete

'Senior chief engineer, Division of Fire Technology, National Testing Institute, Boris, Sweden.

Copyright by ASTM Int'l (all rights reserved); Wed Mar 14 22:28:29 EDT 2018145
Downloaded/printed by
University of Queensland (University of Queensland) pursuant to License Agreement. No further reproductions authorized.
Copyright*' 1985 b y A S T M International www.astm.org
 146 FIRE SAFETY: SCIENCE AND ENGINEERING

Nomenclature

A Opening area
a Thermal diffusivity
At Surrounding area
CQ, C ] , C2 Constants
c Specific heat capacity
h Opening height
I Rate of heat
k Conductivity
q Heat flux rate
T Temperature
t Time
t* Modified time
X Depth
Y ATh/AtZ-ikpc

Greek Letters

ai Flow constant
"2 Burning efficiency
Compartment time factor
r 72 = r
7
e Effective emissivity
1 Temperature rise ratio
Temperature rise T — TQ
e Density
p
a Stefan-Boltzmann constant
T( Time constant

Subscripts
A Standard fire compartment
a Thermal diffusivity
C By combustion
c Normal weight concrete
d Duration
/ Fire
L By convection
i Thermal inertia
R By radiation
W Walls
x.y Depth
0 Ambient, initial

Copyright by ASTM Int'l (all rights reserved); Wed Mar 14 22:28:29 EDT 2018
Downloaded/printed by
University of Queensland (University of Queensland) pursuant to License Agreement. No further reproductions authorized.
 WICKSTROM ON STANDARD FIRE CURVE 147

Fire engineering design of structures and structural elements are in most

cases characterized by a procedure based on a classification system and an asso-
ciated standardized fire test, ISO 834, with fixed heating conditions. This test
exposure simulates a fully developed fire of prescribed duration; in general,
only the heating phase is considered. However, in many countries there is a
growing interest in computing the fire exposure from real or natural conditions.
Thus, for several years now the Swedish building code has allowed for an engi-
neered design based on heat and mass balance analysis of fire compartments.
The fire compartment is taken to be a well-stirred reactor, with the following
major assumptions:
1. The fire is ventilation-controlled.
2. The compartment is ventilated by natural convection.
3. The gas temperature is uniform in the fire compartment.
4. The fuel bums wholly inside the fire compartment.
All fires are, of course, not ventilation-controlled, and particularly in large
spaces local fires may have a very severe impact on structural elements. Fires in
smaller spaces severe enough to threaten structural elements tend, however, to
be ventilation-controlled. Using this theory, the fire temperature in the heat-
ing phase is derived with a Laplace transform technique and compared with
the ISO 834 standard time-temperature curve. It will be shown that ventilation
and wall properties may be taken into account by altering the time scale, and
consequently the development of fire temperatures under various conditions
may be expressed in a single diagram. For a certain set of input parameters,
this temperature diagram nearly coincides with the ISO 834 standard curve for
the first several hours of a fire. It therefore is suggested that this well-estab-
lished temperature relation also may be employed for fire engineering design
when natural fires are assumed. The thermal response of concrete structures
exposed to the derived fire temperatures then may be given in a very compact
way. The thermal properties of the concrete structure and the fire compart-
ment in general are different. The mathematical properties of the time-tem-
perature function suggested here make it possible to plot surface temperature
and interior temperature in single diagrams. Fire characteristics and thermal
properties of the concrete are considered by modifying the time scales in suit-
able ways. Thus a simple method for assessing temperature in concrete with
various thermal properties exposed to natural fires is obtained.
The work extends the earlier concept [1] of modified time to describe nat-
ural fires for insulated steel structures. In that paper, a general natural fire
curve was defined based on the ventilation-controlled fire description of Mag-
nusson and Thelandersson [2,3].
A different approach to rational fire engineering design has recently been
introduced by Harmathy. In a series of papers (see, for example, Refs 4 and 5),
the severity of fires is characterized by a single parameter named the normal-
ized heat load, which is explicitly related to fire load, thermal inertia of com-

Copyright by ASTM Int'l (all rights reserved); Wed Mar 14 22:28:29 EDT 2018
Downloaded/printed by
University of Queensland (University of Queensland) pursuant to License Agreement. No further reproductions authorized.
 148 FIRE SAFETY: SCIENCE AND ENGINEERING

partment boundaries, and ventilation properties. The relation of this ap-

proach to standard furnace testing has been explored as well [6].

Temperature Analysis of Postflashover Fires

The theory of the fully developed or the postflashover compartment fire is
summarized in this section. Such fires have been analyzed amply by many
authors [2,3,7-9] and therefore will not be discussed in detail here.
The heat balance of a compartment fire as shown in Fig. 1 may be written as
IC = IL+IW + IR (1)
where
Ic = the heat release rate by combustion,
Ii = the heat loss rate by convection out of the compartment openings,
lyv = the heat transfer to the compartment boundaries, and
IR = the radiation rate out of the openings.
Air and combustion products flow in and out of the compartment driven by
buoyancy at a rate of aj^lV^, where aj is a flow constant and AyJlt the ventila-
tion parameter; for a single opening compartment A and h are the opening
area and height, respectively.
For a ventilation-limited fire, the combustion rate inside the compartment
is proportional to the available oxygen in the incoming air

Ic = Oil ctjAyfTi (2)

where
a2 = a constant describing the burning efficiency and the combustion en-
ergy developed per unit mass of air.

FIG. 1—Heat flow balance of afire compartment.

Copyright by ASTM Int'l (all rights reserved); Wed Mar 14 22:28:29 EDT 2018
Downloaded/printed by
University of Queensland (University of Queensland) pursuant to License Agreement. No further reproductions authorized.
 WICKSTROM ON STANDARD FIRE CURVE 149

The convection loss term is

Ii = caiA^hef (3)
where
c = the specific heat of air, and
Qf — the temperature rise of the fire above ambient conditions.
The wall loss rate is
Iw = Atqw (4)

where
Ai = the area of the compartment boundaries, and
qw — the heat flux rate.
For typical fire durations, the thermal thickness of most walls is so great that
they may be considered semi-infinite. Thus, q^h proportional to the thermal
inertia yfjkpc) of the wall material. In this simple model constant effective val-
ues of the thermal inertia must be chosen which incorporate effects of material
properties varying with temperature as well as effects of thin, exposed layers
where the conditions on the backside influence the thermal response of the ex-
posed surface. In most cases this influence is believed to be very small. The in-
fluence of the heat transfer resistance is small, and for the sake of simplicity it
is ignored in this study, that is, the wall surface temperature and the fire tem-
perature are assumed equal. The heat flux at the surface for a unit tempera-
ture rise is ^kpc/irt), and q^ is obtained by convolution between this
response function and the rate of temperature rise of the fire, 0y.
The radiation loss through the openings is

4 = ^ 9s (5)
where
qr is approximated as aTj, where a is the Stef an-Boltzmann constant and Tf
the absolute fire temperature.
Inserting Eqs 2 through 5 into Eq 1, followed by rearranging the heat bal-
ance equation, yields

C, = C, Qyit) + ^^ [ ^ 0 / (, - | ) d ^ + ^ (6)
'0

where
C\ = cai,
Ci = aiQ!2, and
y = {A^h/AtV^ihpc).

Copyright by ASTM Int'l (all rights reserved); Wed Mar 14 22:28:29 EDT 2018
Downloaded/printed by
University of Queensland (University of Queensland) pursuant to License Agreement. No further reproductions authorized.
 150 FIRE SAFETY: SCIENCE AND ENGINEERING

0y {t — 0 is the time derivative of the fire temperature at the time {t — 0. The

numerator, A'iJi/Af, is recognized as the opening factor [3]. The concept of
modified time, t* = yh, is now introduced where the dimensionless fire com-
partment time factor F = 7^ is defined by 7 = Y/Yy^, where Y^ is a constant
characterizing a fire compartment with standard ventilation and wall proper-
ties

F4 = (0.04)/(1160) = 34.5 X lO-*- m^^^Ks-^^^ W~^ (7)

After changing the time variables, Eq 6 yields

Since the height, h, of the compartment openings varies in general only

within a small range and the radiation term is small, the influence of h on fire
temperature is small and may be ignored. An examination of Eq 8 then reveals
that the fire temperature is a universal function of modified time, t*, only.
Thus when the time-temperature curves reported in Ref 2 were plotted ver-
sus modified time, a single curve was approximately obtained for the heating
phase [1]; the curve was named the general natural fire curve.

Analytical Solution for the Ventilation-controlled Compartment Fire

If the radiation loss /«is neglected and material properties are assumed con-
stant, Eq 6 may be solved by employing Laplace theory and the temperature
rise may be given in a closed analytical form^ as

Qy = (C2/C1) [1 - exp it/Tf) erfc yRt/Tj)] (9)

where

Thus the fire temperature development may be expressed by a maximum tem-

perature C2/C1 and a time constant Tf. If the radiation losses are considered
another parameter, "fh must in principle be introduced and an analytical solu-
tion may not be derived. The window height varies, however, within a very
small range, and its influence may be included in the choice of temperature
function. For Tf — 0.33 h and C2/C1 = 1325°C Eq 9 approximately coincides
for several hours with the standard fire curve according to ISO 834, that is,
Qj = 345 log (480t + 1), where the time t is in hours. In Fig. 2, these two rela-

^Personal communication with Goren Sandberg, Lund Institute of Technology, Lund, Sweden.

Copyright by ASTM Int'l (all rights reserved); Wed Mar 14 22:28:29 EDT 2018
Downloaded/printed by
University of Queensland (University of Queensland) pursuant to License Agreement. No further reproductions authorized.
 WICKSTROM ON STANDARD FIRE CURVE 151

ef[»ci

FIG. 2—Fire temperature curves.

tions are plotted as well as the general natural fire curve as defined in Ref 1 and
an exponential approximation to the standard curve specified in the former
Swedish building code, SBN 1975 [10]. During the first several hours, all four
curves differ very little.

The Standard Fiie Curve to Express Natural Fires

As the ISO curve is well-known, it would be advantageous if it could be used
for design purposes together with the concept of modified time. It is, however,
impossible as it goes to unrealistically high temperatures for fire durations that
very well may occur in the modified time scale. Therefore, it is suggested that
the exponential approximation as given in the former Swedish building code
[10] may be employed, that is
3
(10)

where 5, and ft- are given in Table 1.

Copyright by ASTM Int'l (all rights reserved); Wed Mar 14 22:28:29 EDT 2018
Downloaded/printed by
University of Queensland (University of Queensland) pursuant to License Agreement. No further reproductions authorized.
 152 FIRE SAFETY: SCIENCE AND ENGINEERING

TABLE 1--Constants in the exponential expression {UqW) of the

ISO 834 standard curve [10].

j 0 1 2 3

B(°C) 1325 -430 -270 -625

3(A-l) 0 0.2 1.7 19

Temperature in Fire-Exposed Concrete Structures

In this section, a technique for estimating the temperature response of ex-
ample concrete structures will be shown. The structure is exposed to a given
fire; its thermal properties differ in the general case from those characterizing
the compartment walls. The design curves given are obtained by plotting nu-
merically calculated temperatures versus time scaled by suitable parameter
groups. Any rigid mathematical derivations were not carried out.
In a first step, only semi-infinite, one-dimensional structures will be consid-
ered. The time variations of the ratios r]w = Q^/Qj and r;^ = Qjc^Qw will be
shown, where 9y, Q^,, and 6^ are the temperature rises of the fire, the exposed
surface, and a point inside the structure, respectively. Various fires character-
ized by the fire compartment time factor and various material properties of the
exposed structure are taken into account.
A computer program, TASEF-2 [11,12], based on the finite-element
method was employed to calculate the thermal response. The heat transfer
rate per unit area by radiation from the fire to the structure is assumed equal to
e a {Tf — rtv), where e is an effective emissivity, cr the Stefan-Boltzmann con-
stant, and Tf and Tw the absolute temperatures of the fire and the surface, re-
spectively. The convection to the walls is assumed equal to h{Tf — Ty^^), where
h is the heat transfer coefficient. In all calculations reported here, e = 0.7 and
ft = 25Wm-2K:-i.
The plots shown below are based on computer analyses of one-dimensional,
semi-infinite structures where the assumed material properties, as given in Ta-
ble 2, simulate normal-weight concrete and light-weight concrete, respectively.

TABLE 2—Assumed material properties of normal-weight concrete and light-weight concrete [13].

Normal-Weight Light-Weight
Concrete Concrete

Density, p (kg m~^) 2400 500

Specific heat capacity, e (J kg~' K~') 1000 1000
Thermal conductivity, k (W m ~ ' K " ' ) 1.0" 0.3
Thermal inertia, -fikpc) (W m~^ s~'^^) 1550 387
Thermal diftusivity, a X 10'
(m^s-') 417 600
(m^h"') 0.116 0.167

"Variations in thermal conductivity prescribed by Ref 13 are shown in Fig. 3.

Copyright by ASTM Int'l (all rights reserved); Wed Mar 14 22:28:29 EDT 2018
Downloaded/printed by
University of Queensland (University of Queensland) pursuant to License Agreement. No further reproductions authorized.
 WICKSTROM ON STANDARD FIRE CURVE 153

The values are taken from the Swedish Building Code [13]. In some cases the
thermal conductivity of concrete is assumed constant with the purpose of test-
ing the theory and creating the graphs for material with constant material
properties in general.
Figure 3 shows that for a given material the ratio ^]^, — Q^/Qf may approxi-
mately be given as a function of yt only. Of rjjy is plotted versus tw — (7/7,) t as
in Fig. 4, the material properties of the actual structure may be included in one
diagram as well. 7, is the ratio between the thermal inertia of the material of
the actual structure and that of normal-weight concrete (see Table 2)

7,- = nkpc)/r{kpcX (11)

The surface temperature rise may now be estimated as

Qw = i)wQf (12)

1.0
Light-weight , ^ *' * 0*9* « * * • * *
concrete , « ^»v*'*

% Normal weight
concrete
0.9

0.8
• Y=10
« 7 = 2.0

TlW =
Of
0.7

0.6 _, , , , 1
0.1 1.0 2,0 3.0 VO YHhl

FIG. 3—The ratio •^ between the surface temperature rise Qy/ and the fire temperature rise 9f
plotted versus 7 ! The thermal conductivity of normal weight concrete varied according to Fig. 3.

Copyright by ASTM Int'l (all rights reserved); Wed Mar 14 22:28:29 EDT 2018
Downloaded/printed by
University of Queensland (University of Queensland) pursuant to License Agreement. No further reproductions authorized.
 154 FIRE SAFETY: SCIENCE A N D ENGINEERING

•nw.-n

FIG. 4—Upper curve and upper horizontal axis: the ratio 1}^ versus t^ = ly/yjt. Lower curve
and lower horizontal axis: the ratio -q,, versus fi^ = y^ t/x^.

The temperatures obtained with this approximation differ by less than a few
percent from the values obtained with TASEF-2.
Inside a structure the temperature rise 9^. at a distance x from the surface
may also be obtained from Fig. 4 where the ratio rj^ = QJQ-w is plotted versus
"x = Tfli/jc^. The coefficient 7„ is the ratio between the thermal diffusivity of
the actual structure and of normal-weight concrete (see Table 2)

7« = a/a. (13)

When assuming constant material properties, -q^ lies within 1 or 2% from

the full line. Many combinations were checked with x equals 0.04 m and 0.10
m and t up to four h. For concrete structures with a varying conductivity, ac-
cording to Fig. 5, the results follow a different graph (dashed line in Fig. 4).
Still, the accuracy is within a few percent. Thus, the temperature at an internal
point may be calculated as

e, =,,, ew (14)

Copyright by ASTM Int'l (all rights reserved); Wed Mar 14 22:28:29 EDT 2018
Downloaded/printed by
University of Queensland (University of Queensland) pursuant to License Agreement. No further reproductions authorized.
 WICKSTROM ON STANDARD FIRE CURVE 155

Thermal conductivity, k
[Wm-iKl

\
^^
1 0-

0-
0 500 1000 Temp°C

FIG. 5—Thermal conductivity of normal weight concrete as a function of temperature [1^ •

and by inserting Eq 12

©I - ^x Vw 9, (15)
'/
Observe that the approximations as outlined here are particularly simple to use
when analyzing normal weight concrete structures exposed to the standard fire
curve, that is y, 7,, and y^ are equal to unity.

Other Geometric^ Configurations

So far only thermally semi-infinite, one-dimensional bodies have been stud-
ied; a wall is thermally semi-infinite if the second boundary has negligible in-
fluence on the temperature development in the area considered. This condi-
tion is satisfied when the penetration depth 8 is less than the thickness of the
wall plus the distance back to the point to be analyzed (see Fig. 6). The wall

FIG. 6—When estimating the temperature at the distance xfrom the exposed surface, the wall
may be assumed infinitely thick if the penetration depth & > (2d — x), where 6 = J.6V7atj.

Copyright by ASTM Int'l (all rights reserved); Wed Mar 14 22:28:29 EDT 2018
Downloaded/printed by
University of Queensland (University of Queensland) pursuant to License Agreement. No further reproductions authorized.
 156 FIRE SAFETY: SCIENCE AND ENGINEERING

may be assumed semi-infinitely thick if 8 is less than 2d — x. The penetration

depth may be estimated as
8 = 3.6 yTiat) (16)
In the case of a wall exposed to a fire from both sides, the temperature re-
sponse from each side may be superimposed and T;^ is approximated by rj]f*' +
ij"**"*, where the distance from the two sides is considered.
Two-dimensional square corners may as well be analyzed by superposition
with contributions from the two perpendicular boundaries. The temperature
right in the comer edge is assumed equal to the fire temperature, and the
formula

©xy = [VW iVx + Vy- ^Vx ly) + Vx Vy] ©/ (17)

is derived where 0 ^ is the temperature rise at the distances otx a n d j to the two
perpendicular surfaces, respectively. The values of rj^ and the corresponding rjy
are obtained from Fig. 4 as just described for one-dimensional configurations.

Average Temperatura of Reinforcement Bars at Different Distances from

Exposed Surface
In some cases it is of interest to calculate an average temperature of a group
of reinforcement bars embedded at various depths. This of course could be
carried out by calculating the temperature of each bar of interest, but a more
convenient way is to calculate an effective depth x and then calculate the tem-
perature at that point. Figure 4 shows that the temperature rise is linear with x
in a logarithmic scale.
As the equation log{t/x^) = [log(t/xf) + logit/xj)]/! yields x = {xi ^2)'^^,
the effective distance x is in the general case the geometric mean value

x = (xiX2X3 . . . ; c „ ) i / « (18)

where x, is the distance to the surface of each of the reinforcement bars. In the
case of bars of various areas A;, the weighted effective distance is obtained as

x = ixi'xi^xj' ...x^'-y/^ (19)

where A is the sum of A,-.

Observe that it is an approximation on the unsafe side to assume x as the
arithmetic mean value. The arithmetic mean value of 2 and 8 is, for instance,
5, while the corresponding geometric mean value is not more than 4.

Maximum Temperatures Reached During a Complete Compartment Fire

When designing a structure exposed to natural fire conditions, the cooling
phase must in general be included. The heating phase duration is dependent

Copyright by ASTM Int'l (all rights reserved); Wed Mar 14 22:28:29 EDT 2018
Downloaded/printed by
University of Queensland (University of Queensland) pursuant to License Agreement. No further reproductions authorized.
 WICKSTROM ON STANDARD FIRE CURVE 157

on the amount and the characteristics of the fuel load and the ventilation con-
ditions. The assumption that the fire is ventilation-controlled and that the fuel
bums wholly inside the compartment implies that the heating phase duration
is proportional to the fuel load density divided by the ventilation factor; al-
though very crude, such an estimate is suggested, for example, in the Swedish
building code. The fuel load density is defined as the total heat of combustible
material per unit internal surface area of the compartment. While during the
heating phase the temperature may be estimated as just described, a good esti-
mate during the cooling phase is more difficult to express in a general way.
In the international standard ISO 834, the cooling phase temperature is as-
sumed to decline linearly with time; the rates vary with the fire durations.
These decay curves are rough approximations of natural fires [2], and, as the
fire temperature in the cooling phase is hard to estimate on a sound physical
basis, it is here suggested that the linear time temperature relation as described
in ISO 834 should be employed for the type of analysis introduced here as well.
The maximum temperature at a point in a structure exposed to a complete
fire history is not reached until some time has elapsed into the cooling phase.
Figure 7 shows an example of how the maximum temperature rise may be ob-
tained in a structure exposed to a fire with a heating phase duration t^ = 1 h,
followed by a cooling phase according to ISO 834. An unlimited fire is
simulated and G^ax is calculated at the time t = t^ + At where At depends on
the depth of the point considered. By experience from several computer
analyses, the following approximate estimation is obtained

Ai = ;uxVa (20)

With fjL = 0.34, this approximation procedure predicts the maximum temper-
ature with an accuracy of about 5% on the condition that the predicted tem-
perature is over 400°C and the fire duration is less than 1 h.

Examples
Estimate the temperature rise 4 cm into a semi-infinitely thick concrete wall
after a fire duration, t^ = 1 h. Assume standard material properties as given in
Table 2 with variable thermal conductivity as shown in Fig. 5 (7,- = Ja — 1),
and a fire characterized by a fire compartment time factor, F = 4.
Start by calculating the modified fire duration ^* = 4 h and get 6y =
1130°C from Fig. 2; then calculate tyy = y tj = 2h and /j.^ = t^/x^ = 625 h
m~^ and get from Fig. 4 rj^ = 0.97 and r/^ = 0.345, respectively. Now the
temperature rise is calculated as 9^. = rj^ rjvy Qy = 379°C. A finite-element
analysis of the same example gave 0^. = 380°C.
Then calculate the maximum temperature at the same point if the cooling
phase temperature development is according to ISO 834.
A time increment At = 0.36 h is calculated according to Eq 20. Thus, the ef-
fective time is 1.36 h or in the modified time scale 5.44 h. Now continue the cal-

Copyright by ASTM Int'l (all rights reserved); Wed Mar 14 22:28:29 EDT 2018
Downloaded/printed by
University of Queensland (University of Queensland) pursuant to License Agreement. No further reproductions authorized.
 158 FIRE SAFETY: SCIENCE AND ENGINEERING

ecci

1000-

Limited fire duration

Unlimited fire duration

500-
Concrete temperature
xsOOim

2.0 t[hl

FIG. l^The maximum temperature is estimated by increasing the fire duration and carrying
out the analysis at a fictitious time (t^ + At). The graphs are obtained from finite-element com-
puter calculations [11].

culation as above and get the maximum temperature 0"*" = 446°C. A com-
puter analysis gave Gf ^^ = 446°C.
Finally, calculate the temperature rise in a comer of a concrete structure at a
point X = 4 cm and 3; = 6 cm from the boundaries after a 2-h exposure to the
standard fire curve (r = 1). With the same procedure as above 9^ = 1028°C,
riw = 0.97, rj^ = 0.475, and 17^ = 0.325. Then Eq 17 yields 0 ^ = 649°C. A
corresponding computer analysis gave 0 ^ = 650°C.

Copyright by ASTM Int'l (all rights reserved); Wed Mar 14 22:28:29 EDT 2018
Downloaded/printed by
University of Queensland (University of Queensland) pursuant to License Agreement. No further reproductions authorized.
 WICKSTROM ON STANDARD FIRE CURVE 159

Summaiy
The fully developed compartment fire has been analyzed, and the sugges-
tion is made that the standard fire curve according to ISO 834 may be used as a
basis for describing natural fires. The influence of ventilation condition and
thermal properties of surrounding structures is considered by modifying the
time scale. However, to avoid unrealistically high temperature for modified
fire durations exceeding 10 h, the logarithmic expression of the ISO curve is re-
placed by a sum of exponential terms.
It has also been shown how the temperature in fire-exposed concrete struc-
tures can be estimated by employing simple diagrams, Figs. 2 and 4. Various
fire conditions and material properties may be considered. One- and two-
dimensional structures were studied and the suggested approximation proce-
dures were shown to yield results that only deviated a few percent from the re-
sults obtained by comprehensive computer analyses.

Acknowledgments
Stig Andersson was most helpful with the calculation work of this paper.
His contribution to the design and plotting of the diagrams is also greatly
appreciated.

Referances
[1] Wickstrom, Ulf, Fire Safety Journal, Vol. 4, 1981, 1982, p. 219.
[2] Magnusson, S.-E. and Thelandersson, S.,Acta Polytechnica Scandinavia, Civil Engineering
Building Construction, Series No. 65, 1970; or "Temperature-Time Curves of Complete Pro-
cess of Fire Development," Bulletin 19, Division of Structural Mechanics and Concrete Con-
struction, Lund Institute of Technology, Lund, 1970.
[3] Magnusson, S.-E. and Thelandersson, S., "Comments on Rate of Gas Flow and Rate of Burn-
ing for Fires in Enclosures," Bulletin 19, Division of Structural Mechanics and Concrete
Construction, Lund Institute of Technology, Lund, 1971.
[4] Harmathy, T. Z. and Mehaffey, J. R., Fire and Materials, Vol. 7, 1983, p. 49.
[5] Harmathy, T. Z., Fire and Materials, Vol. 4, 1980, p. 71.
[6] Harmathy, T. Z.,Fire and Materials, Vol. 5, 1981, p. 112.
[7\ Kawagoe, K. and Sekine, T., "Estimation of Fire Temperature-Time Curve in Rooms," BRI
Occasional Report 11, Building Research Institute, Tokyo, 1963.
[8\ Odeen, K., "Theoretical Study of Fire Characteristics in Enclosed Spaces," Bulletin 10, Di-
vision of Building Construction, Royal Institute of Technology, Stockholm, 1963.
[9] Babrauskas, V., "Fire Endurance in Buildings," Report No. UCB FRO 76-16, Fire Re-
search Group, Department of Civil Engineering, University of California, Berkeley, 1977.
[10\ The Swedish Building Code, SBN 1975, The National Swedish Board of Physical Planning
and Building, Stockholm.
[//] Wickstrom, Ulf, "TASEF-2—A Computer Program for Temperature Analysis of Structures
Exposed to Fire," Report No. 79-2, Department of Structural Mechanics, Lund Institute of
Technology, Lund, 1979.
[12] Paulsson, Mats, "PRETEMP—Preprocessor for Finite Element Analysis of Non-linear
Transient Heat Conduction," LUTVDG/(TVSM-7011)/l-79/(1983). Division of Structural
Mechanics, Lund Institute of Technology, Lund, 1983.
[13] "Rules for Approval of Loadbearing Capacity of Building Elements Subjected to Fire," PFS
1984:1, The National Swedish Board of Physical Planning and Building, Stockholm, 1984.

Copyright by ASTM Int'l (all rights reserved); Wed Mar 14 22:28:29 EDT 2018
Downloaded/printed by
University of Queensland (University of Queensland) pursuant to License Agreement. No further reproductions authorized.