Example to EN 1991 Part 1-2: Compartment fire

P. Schaumann, T. Trautmann
University of Hannover – Institute for Steel Construction, Hannover, Germany

1 TASK

The gas temperature of a fully engulfed fire in an office has to be determined.

The room of the “Simulated Office” test of the Cardington building is chosen
for this analysis. The measured temperatures during the fully engulfed fire are
shown in Figure 3, so the calculation can be compared with these results.
A natural fire model is chosen for the calculation of the gas temperature.
For fires with a flash-over, the method of the compartment fires can be used.
A simple calculation method for a parametric temperature-time curve is given
in Annex A of EN 1991-1-2.

Figure 1. Cardington building (left) and the office of the “Simulated Office” test (right)

Floor area: Af = 135 m²

Total area of enclosures At = 474 m²
Total area of vertical openings: Av = 27 m²
Vertical opening factor: αv = 0.2
Horizontal opening factor: αh = 0.0
Height: H = 4.0 m
Average window height: heq = 1.8 m (assumption)
Lightweight concrete: ρ = 1900 kg/m²
c = 840 J/kgK
λ = 1.0 W/mK
Fire growth rate medium
2 DETERMINATION OF FIRE LOAD DENSITY EN 1991-1-2

For the determination of the fire load density the Annex E of EN 1991-1-2
offers a calculation model. The design value of the load density may either be
given from a national fire load classification of occupancies and/or specific for
an individual project by performing a fire load evaluation.
At this example, the second method is chosen.
q f ,d = q f ,k ⋅ m ⋅ δ q1 ⋅ δ q 2 ⋅ δ n Annex E.1

where:
m the combustion factor
δq1 the factor considering the danger of fire activation by size of the
compartment
δq2 the factor considering the fire activation risk due to the type of occupancy
δn the factor considering the different active fire fighting measures

The fire load consisted of 20 % plastics, 11 % paper and 69 % wood, so it

consisted mainly of cellulosic material. Therefore the combustion factor is:
m = 0.8
The factor δq1 considers the danger of fire activation by size of the
compartment, as given in Table 1.

Table 1. Fire activation risk due to the size of the compartment (see EN 1991-1-2,
Table E.1)
Compartment floor area Af [m²]
≤ 25 ≤ 250 ≤ 2500 ≤ 5000 ≤ 10,000
Danger of fire
1.10 1.50 1.90 2.00 2.13
activation δq1

δq1 = 1.5
A factor δq2 considers the fire activation risk due to the type of occupancy, as
given in Table 2.

Table 2. Fire activation risk due to the type of occupancy (see EN 1991-1-2,
Table E.1)
Danger of fire
Examples of occupancies
activation δq2
0.78 artgallery, museum, swimming pool
1.00 offices, residence, hotel, paper industry
1.22 manufactory for machinery & engines
1.44 chemical laboratory, painting workshop
1.66 manufactory for fireworks or paints

δq2 = 1.0
The factor taking the different active fire fighting measures into account is
calculated to:
10
δ n = ∏ δ ni
i =1

The factors δni are given in Table 3.

Table 3. Factors δni (see EN 1991-1-2, Table E.2)
δni function of active fire fighting measures
Automatic fire Automatic water extinguishing
δn1 0.61
suppression system
0 1.0
Independent water supplies δn2 1 0.87
2 0.7
Automatic fire
δn3 By heat or 0.87
detection Automatic fire detection & alarm
δn4 by smoke 0.73
Automatic alarm transmission to
δn5 0.87
fire brigade
Manual fire
Work Fire Brigade δn6 0.61
suppression
Off Site Fire Brigade δn7 0.78
0.9 or 1.0
Safe access routes δn8
or 1.5
Fire fighting devices δn9 1.0 or 1.5
Smoke exhaust system δn10 1.0 or 1.5

δ n = 1.0 ⋅ 0.73 ⋅ 0.87 ⋅ 0.78 ⋅ 1.0 ⋅ 1.0 ⋅ 1.0 = 0.50

For calculating the characteristic fire load, the characteristic fire load has to be
determined. It is defined as:
Q fi ,k = ∑M k ,i ⋅ H ui ⋅ψ i Annex E.2

where:
Mk,i the amount of combustible material [kg]
Hui the net calorific value [MJ/kg], see EN 1991-1-2, Table E.3
ψi the optional factor for assessing protected fire loads

The total fire loading was equivalent to 46 kg wood/m², so the characteristic

fire load is:
Q fi ,k = (135 ⋅ 46 ) ⋅ 17.5 ⋅ 1.0 = 108,675 MJ
The characteristic fire load density is determined to:
q f ,k = Q fi ,k A f = 108,675 135 = 805 MJ/m²
The design value of the fire load density is calculated to:
q f ,d = 805 ⋅ 0.8 ⋅ 1.5 ⋅ 1.0 ⋅ 0.5
= 483.0 MJ/m²

3 CALCULATION OF THE PARAMETRIC TEMPERATURE-TIME Annex A

CURVE

It has to be determined if the fully engulfed fire is fuel or ventilation

controlled. For this, the opening factor and the design value of the fire load
density related to the total surface are needed.
≥ 0.02
O = heq ⋅ Av At = 1.8 ⋅ 27 474 = 0.076 m1 2 
≤ 0.2
and
qt ,d = q f ,d ⋅ A f At = 483.0 ⋅ 135 474 = 137.6 MJ m 2
The determination, if the fire is fuel or ventilation controlled is:
0.2 ⋅ 10−3 ⋅ qt ,d O = 0.2 ⋅ 10−3 ⋅ 137.6 0.076 = 0.362 h > tlim = 0.333 h

⇒ The fire is ventilation controlled

For calculation of the temperature-time curves for the heating and the cooling
phase, the b factor is needed. This factor considers the thermal absorptivity for
the boundary of enclosure. The density, the specific heat and the thermal
conductivity of the boundary may be taken at ambient temperature. The floor,
the slab and the walls are made of lightweight concrete
J ≥ 100
b = ρ ⋅ c ⋅ λ = 1900 ⋅ 840 ⋅ 1.0 = 1263.3 
2 12
m s K ≤ 2200
The temperature-time curve in the heating phase is given by:

(
θ g = 20 + 1325 ⋅ 1 − 0.324 ⋅ e −0.2⋅t* − 0.204 ⋅ e −1.7⋅t* − 0.472 ⋅ e−19⋅t* )
Because the fire is ventilation controlled, the time t* is calculated to:
t* = t ⋅ Γ
where:

Γ=
( O b )2
=
( 0.076 1263.3)
2

= 3.04
( 0.04 1160 )2 ( 0.04 1160 )2
Now the heating phase can be calculated:

(
θ g = 20 + 1325 ⋅ 1 − 0.324 ⋅ e −0.2⋅( 3.04⋅t ) − 0.204 ⋅ e −1.7⋅( 3.04⋅t ) − 0.472 ⋅ e −19⋅( 3.04⋅t ) )
For calculation of the cooling phase, the maximum temperature is needed.

(
θ max = 20 + 1325 ⋅ 1 − 0.324 ⋅ e −0.2⋅t*max − 0.204 ⋅ e−1.7⋅t*max − 0.427 ⋅ e −19⋅t*max )
where:
t *max = tmax ⋅ Γ

The time tmax is determined as below, where tlim is given in Table 4.

0.2 ⋅ 10 ⋅ qt ,d O = 0.2 ⋅ 10 ⋅ 137.6 0.076 = 0.363 h

−3 −3
tmax = max 
tlim = 0.333 h

Table 4. Time tlim for different fire growth rates

Slow fire growth rate Medium fire growth rate Fast fire growth rate
tlim [h] 0.417 0.333 0.250

So t*max is calculated to:

t *max = 0.363 ⋅ 3.04 = 1.10 h
The maximum temperature is calculated to:

(
θ max = 20 + 1325 ⋅ 1 − 0.324 ⋅ e −0.2⋅1.10 − 0.204 ⋅ e −1.7⋅1.10 − 0.427 ⋅ e−19⋅1.10 )
= 958.8 °C
During the cooling phase, t* and t*max are calculated to:
t* = t ⋅ Γ = t ⋅ 3.04 [h]

( )
t *max = 0.2 ⋅ 10−3 ⋅ qt ,d O ⋅ Γ = 1.10h

The temperature-time curve in the cooling phase for 0.5 ≤ t*max ≤ 2.0
is given by:
θ g = θ max − 250 ⋅ ( 3 − t *max ) ⋅ ( t * −t *max ⋅x )
= 958.8 − 250 ⋅ ( 3 − 1.10 ) ⋅ ( t ⋅ 3.04 − 1.10 ⋅ 1.0 )
where:
tmax > tlim x = 1.0

Combination of the heating and cooling curves leads to the parametric

temperature-time curve shown in Figure 2.

1600

parametric temperature-time curve

1400
Heating phase
1200
Cooling phase
temperature [°C]

1000

800

600

400

200

0
0 10 20 30 40 50 60 70 80
time t [min]
Figure 2. Gas temperature of the office calculated by using the parametric
temperature-time curve of the office
4 COMPARISON BETWEEN CALCULATION AND FIRE TEST

To compare the calculation with the measured temperatures in the test, the
factors δ1, δ2 and δni for calculation of the fire load density have to be set to
1.0 (see Figure 3).

Figure 3. Comparison of measured and calculated temperature-time curves

REFERENCES

EN 1991, Eurocode 1:Actions on structures – Part 1-2: General actions – Actions on

structures exposed to fire, Brussels: CEN, November 2002
The Behaviour of multi-storey steel framed buildings in fire, Moorgate: British Steel
plc, Swinden Technology Centre, 1998
Valorisation Project: Natural Fire Safety Concept, Sponsored by ECSC, June 2001
Example to EN 1991 Part 1-2: Localised fire

P. Schaumann, T. Trautmann
University of Hannover – Institute for Steel Construction, Hannover, Germany

1 TASK

The steel temperature of a beam has to be determined. It is part of an

underground car park below the shopping mall Auchan in Luxembourg. The
beams of the car park are accomplished without any use of fire protection
material. The most severe fire scenario is a burning car in the middle of the
beam (see Figure 1).
For getting the steel temperature, the natural fire model of a localised fire is
used.

Figure 1. Underground car park of the shopping mall Auchan

Figure 2. Static system and cross-section of the beam

 Diameter of the fire: D = 2.0 m
Vertical distance between fire source
and ceiling: H = 2.7 m
Horizontal distance between beam
and flame axis: r = 0.0 m
Emissivity of the fire: εf = 1.0
Configuration factor: Φ = 1.0
Stephan Boltzmann constant: σ = 5.67 · 10-8 W/m2K4
Coefficient of the heat transfer: αc = 25.0 W/m²K

Steel profile: IPE 550

Section factor: Am/V = 140 1/m
Unit mass: ρa = 7850 kg/m³
Surface emissivity: εm = 0.7
Correction factor: ksh = 1.0

2 RATE OF HEAT RELEASE ECSC Project

The rate of heat release is normally determined by using the EN 1991-1-2

Section E.4. For dimensioning the beams at this car park, the rate of heat
release for one car is taken from an ECSC project called "Development of
design rules for steel structures subjected to natural fires in CLOSED CAR
PARKS" (see Figure 3).

Figure 3. Rate of heat release of one car

3 CALCULATION OF THE STEEL TEMPERATURES EN 1991-1-2

3.1 Calculation of the flame length Annex C

First of all, the flame length has to be determined.
L f = −1.02 ⋅ D + 0.0148 ⋅ Q 2 5 = −2.04 + 0.0148 ⋅ Q 2 5
A plot of this function with the values of Figure 3 is shown in Figure 4. With a
ceiling height of 2.80 m, the flame is impacting the ceiling at a time from 16.9
min to 35.3 min (see Figure 4).

Figure 4. Flame length of the localised fire

It is important to know, if the flame is impacting the ceiling or not, because

different calculation methods for the calculation of the net heat flux are used
for these two cases (see Figure 5).

Figure 5. Flame models: Flame is not impacting the ceiling (A); Flame is impacting
the ceiling (B)

3.2 Calculation of the net heat flux

3.2.1 1st case: The flame is not impacting the ceiling

The net heat flux is calculated according to Section 3.1 of EN 1991-1-2.

h&net = α c ⋅ θ ( z ) − θ m + Φ ⋅ ε m ⋅ ε f ⋅ σ ⋅  θ( z ) + 273 − (θ m + 273) 

4
( ) ( 
) 4


Section 3.1
= 25.0 ⋅ θ( z ) − θ m + 3.969 ⋅ 10−8 ⋅  θ( z ) + 273 − (θ m + 273) 
4
( ) ( 
) 4


The gas temperature is calculated to: Annex C
23 −5 3
θ( z ) = 20 + 0.25 ⋅ ( 0.8 ⋅ Q ) ⋅ ( z − z0 ) ≤ 900 °C
−5 3
= 20 + 0.25 ⋅ ( 0.8 ⋅ Q )
23
(
⋅ 4.74 − 0.0052 ⋅ Q 2 5 ) ≤ 900 °C

where:
z is the height along the flame axis (2.7 m)
z0 is the virtual origin of the axis [m]
z0 = −1.02 ⋅ D + 0.0052 ⋅ Q 2 5 = −2.04 + 0.0052 ⋅ Q 2 5

3.2.2 2nd case: The flame is impacting the ceiling

The net heat flux, if the flame is impacting the ceiling, is given by:
4
(
h&net = h&− α c ⋅ (θ m − 20 ) − Φ ⋅ ε m ⋅ ε f ⋅ σ ⋅ (θ m + 273) − ( 293)
4
)
⋅ ( (θ − ( 293) )
4 4
= h&− 25.0 ⋅ (θ m − 20 ) − 3.969 ⋅ 10−8 m + 273)

The heat flux depends on the parameter y. For different dimensions of y,

different equations for determination of the heat flux have to be used.
if y ≤ 0.30:
h&= 100,000
if 0.30 < y < 1.0:
h&= 136,300 − 121,000 ⋅ y
if y ≥ 1.0:
h&= 15,000 ⋅ y −3.7
where:
r + H + z' 2.7 + z '
y= =
Lh + H + z ' Lh + 2.7 + z '
The horizontal flame length is calculated to:

(
Lh = 2.9 ⋅ H ⋅ QH * ( )
0.33
) − H = (7.83 ⋅ (Q H
*
)
0.33
) − 2.7
where:
( )
QH * = Q 1.11 ⋅ 106 ⋅ H 2.5 = Q 1.11 ⋅ 106 ⋅ 2.7 2.5 ( )
The vertical position of the virtual heat source is determined to:
if QD* < 1.0:

((
z ' = 2.4 ⋅ D ⋅ QD* )
25
− QD*( )
23
) = 4.8 ⋅ ((Q D
*
)
25
− QD*( )
23
)
if QD* ≥ 1.0:

(
z ' = 2.4 ⋅ D ⋅ 1.0 − QD*( )
25
) = 4.8 ⋅ (1.0 − (Q ) ) D
*
25

where:
( )
QD* = Q 1.11 ⋅ 106 ⋅ D 2.5 = Q 1.11 ⋅ 106 ⋅ 2.02.5 ( )
3.3 Calculation of the steel temperature-time curve EN 1993-1-2
The specific heat of the steel ca is needed to calculate the steel temperature.
The parameter is given by EN 1993-1-2, Section 3.4.1.2 depending on the
steel temperature.
Section 3.4.1.2

Figure 6. Specific heat of carbon steel (see EN 1993 Part 1-2, Figure 3.4)
Am V & 1.78 ⋅ 10−2 &
θ a ,t = θ m + k sh ⋅ ⋅ hnet ⋅ ∆t = θ m + ⋅ hnet Section 4.2.5.1
ca ⋅ ρ a ca
The steel temperature-time curve is shown in Figure 6. Additionally, the
results of the FEM-analysis done by PROFILARBED are shown for
comparison.

Figure 7. Comparison of the temperature-time curve of the calculation and the FEM-
analysis of PROFILARBED

REFERENCES
EN 1991, Eurocode 1:Actions on structures – Part 1-2: General actions – Actions on
structures exposed to fire, Brussels: CEN, November 2002
EN 1993, Eurocode 3: Design of steel structures – Part 1-2: General rules –
Structural fire design, Brussels: CEN, October 2006
ECSC Project, Development of design rules for steel structures subjected to natural
fires in CLOSED CAR PARKS, CEC agreement 7210-SA/211/318/518/620/933,
Brussels, June 1996