The effect of insulation on charring of timber frame members

Alar Just

,

Joachim Schmid, Jürgen König

SP Trätek

SP Report 2010:30

S

P

T

ec

hn

ic

al

R

es

ea

rc

h

In

st

it

ut

e

of

S

w

ed

en

Abstract

This report describes the effect of insulation materials on the charring of timber members.

Particular attention is paid to new types of glass wool with high maximum service

temperatures. The aim is to verify the protective effect of stone wool and to determine

similar properties for heat-resistant glass wool, intended for use in accordance with EN

1995-1-2.

When cladding has fallen off, traditional glass wool insulation provides no significant

protection against fire in the post-protection phase. However, a new form of glass wool

insulation, suitable for use at high maximum service temperatures, is now available. The

behaviour of this material in fire is closer to that of stone wool than is the behaviour of

traditional glass wool.

This report describes analysis of the effects of insulation to protect against the charring of

timber members, based on full-scale wall tests.

Key words:

Timber frame assemblies, fire design, Heat-resistant glass wool, charring rate

SP Sveriges Tekniska Forskningsinstitut

SP Technical Research Institute of Sweden

SP Report 2010:30

ISBN 978-91-86319-68-7

ISSN 0284-5172

Contents

Abstract

2

Contents

3

Preface

5

Summary

6

1

Introduction

7

1.1

Symbols

8

1.2

Abbreviations

8

2

Tests

9

2.1

General

9

2.2

Test assemblies

10

2.3

Test results

11

3

Insulation

12

3.1

Insulation types

12

3.1.1

Mineral wool

12

3.1.2

Stone wool

12

3.1.3

Glass wool

12

3.1.4

Glass wool with high maximum service temperature

12

3.1.5

Tested insulations

13

3.2

Thermal properties of insulation

13

3.3

Temperature rise behind the insulation

14

3.4

Small-scale electrical furnace vs full-scale gas heated furnace

16

3.5

Design models

17

3.5.1

Design model for floor and wall assemblies completely filled with

stone wool according to EN 1995-1-2:2004

17

3.5.2

Design model for Ultimate by ETH

17

3.5.3

Tests

18

3.6

Influence of joints in the insulation

19

3.7

Temperature profiles on the unexposed side

21

3.8

Effect of air gap

21

4

Analysis of charred cross-sections

22

4.1

Test walls

22

4.1.1

General

22

4.1.2

Determination of properties of charred cross-sections

23

4.2

Charring rate

24

4.2.1

Design in accordance with EN 1995-1-2:2004

24

4.2.2

Cross-section factor k

25

4.2.3

Conversion factor k

29

4.2.4

Online determination of charred cross-sections

33

4.3

Charring on the sides of cross-sections

35

5

Comparison with other test results

37

6

Proposals for fire design

38

6.1

Floor joists and wall studs in assemblies of which the cavities are

completely filled with Ultimate glass wool or with stone wool

38

6.2

Floor joists and wall studs in assemblies of which the cavities are

completely filled with traditional glass wool

40

6.3

Floor joists and wall studs in assemblies of which the cavities are

partially filled with glass wool or stone wool

40

7

Conclusions

41

8

References

42

Annex A. Temperature profiles on the unexposed side

43

Preface

This report is produced as part of the Fire In Timber project at Wood Wisdom net.

The report consists of analysis of the results of full-scale wall fire tests performed in

Tallinn in 2008/2009, to investigate the performance of stone wool and new innovative

glass wool insulation in protecting timber frame assemblies against fire.

This work is partly supported by Saint-Gobain Isover and Estonian Forest Industry

Association.

Summary

This report describes the effect of insulation materials on the charring of timber members.

The aim of the work has been to verify the protective properties of stone wool, and to

determine the corresponding properties of the new heat-resistant glass wool for use in

accordance with EN 1995-1-2.

EN 1995-1-2:2004 divides mineral wool insulation into two different groups, depending

on their different behaviour in a fire situation. These two groups are stone wool and glass

wool. When the cladding has fallen off in fire situation, traditional glass wool insulation

provides no further significant protection against fire in the post-protection phase. Stone

wool has a noticeable protection effect.

A new material is now available on the market – glass wool insulation with a high

maximum service temperature. It is produced from glass, using glass wool technology.

Protective properties and behaviour in fire is more similar to stone wool, so that dividing

mineral wool into stone wool and glass wool is no longer required.

This report describes investigation of the performance of stone wool and the new

1

Introduction

Timber frame assemblies are normally built up of the timber wall studs or floor joists,

with a cladding attached to each side of the timber frames. The cavities may be void or

partially or completely filled with insulation. Since the timber frame is sensitive to fire

exposure, it must be effectively protected against fire.

In the design and optimisation of a timber frame assembly with respect to maximising fire

resistance, there exists a hierarchy of contribution to fire resistance of various layers of

the assembly.

The greatest contribution to fire resistance is provided by the cladding on the fire-exposed

side that is first directly exposed to the fire, both with respect to insulation and failure

(fall-off) of the cladding. In general, it is difficult to compensate for poor fire protection

performance of the first layer by improved fire protection performance of the following

layers.

For the stage before failure of the cladding, both stone wool and glass wool insulation

perform approximately equally. However, once the cladding has fallen off and the

insulation is directly exposed to the fire, traditional glass wool insulation will undergo

decomposition, gradually losing its protecting effect for the timber member by surface

recession. Stone wool insulation, provided that it remains in place, will continue to

protect the sides of the timber member facing the cavity.

For small-sized timber frame members in assemblies with heat-resistant cavity insulation,

charring mainly takes place on the narrow, fire-exposed side. Since there is a considerable

heat flux through the insulation to the sides of the member during the stage after failure of

the lining (provided that the cavity insulation remains in place), the effect of increasing

arris rounding becomes dominant and no consolidation of the charring rate is possible.

The rules of EN 1995-1-2 [1] for insulated constructions were developed based on the

assumption that all stone wool products lead to the same results, depending only on the

material's density. Since products in the past were optimised for other aspects (production

costs etc.), and no standard for verification of the fire resistance of mineral wool in fire

resistance tests exists, it can no longer be assumed that all stone wools are equal.

New glass wool insulation with a high maximum service temperature provides protection

comparable with stone wool insulation. This material is produced using conventional

glass wool technology. Today, there is only one producer of the material – Saint -Gobain

Isover. The product name of the material is Ultimate.

The main aim of this report is to give design rules for the post-protection phase of

wooden floor and wall assemblies insulated with the new innovative glass wool with a

high maximum service temperature.

The second aim of the report is to proof the knowledge of the post-protection behaviour

of timber frame studs and joists insulated by stone wool.

1.1

Symbols

Afi

cross-section area of the charred cross-section;

Arec

cross-section area of the corresponding rectangular cross-section;

b

cross-section width;

dchar

charring depth;

dchar,n

notional charring depth;

dchar,s

charring depth along wide side of cross-section;

h

cross-section height;

hins

thickness of insulation;

hp

thickness of cladding on fire exposed side;

hres

maximum height of residual cross-section;

hrec

height of corresponding rectangular cross-section;

ks

cross-section factor;

kp

protection factor;

k2

insulation factor;

k3,k3a, kp2b

post-protection factor for phase 3a; (k

in [1])

k3b, kp2c

post-protection factor for phase 3b

kn

factor to convert the irregular residual cross-section into a notional

rectangular cross-section;

tch

start time of charring;

tf

failure time of cladding;

tf,ins

failure time of insulation;

Wfi

section modulus of the charred cross-section;

Wrec

section modulus of the corresponding rectangular cross-section;

β

one-dimensional design charring rate;

β

notional charring rate;

λ

effective thermal conductivity

ρ

nominal density of insulation;

1.2

Abbreviations

GtA

Gypsum plasterboard, Type A

GtF

Gypsum plasterboard, Type F

RW

rock wool insulation, stone wool insulation

GW

glass wool

2

Tests

2.1

General

A series of full-scale tests was performed to investigate the post-protection behaviour of

the insulations, see [2] and [3].

The aim of the work was to investigate the post-protection effect of the new innovative

glass wool (“Ultimate”), produced by Saint-Gobain Isover. The reference for the fire

resistance tests was stone wool, of a quality as used for fire tests for development of EN

1995-1-2 models. The work also included verifying the protective properties of stone

wool.

Test wall assemblies were fitted with additional thermocouples on the sides and inside of

timber studs and behind the boards and insulation. Observations were done and photos

taken of both the fire-exposed side and the unexposed side. The tests were also

documented by thermocamera.

Tests were performed for 30 to 60 minutes until the timber had charred to a depth of

60 mm on the narrow side.

Three of six full-scale tests were performed without cladding on the exposed side in order

to reduce the influence of hard predictable failure of the cladding. The test walls

incorporated two to four different insulations for each test. See Figure 1.

Ultimate

glass wool was tested in comparison to stone wool, without cladding and with

cladding. Additionally, one test was performed of a special construction, with an air gap

between facing and insulation in order to see the effect of an air gap.

Small-scale tests with the same insulation materials from the same batch were performed

in the fire laboratory at the Saint-Gobain Isover factory in Lübz. Insulation materials were

produced in the factories and sent to both fire laboratories in order to compare the

different furnaces.

2.2

Test assemblies

Test 2.1

Test 2.2

Test 2.5

Test 2.6

Figure 1 – Test assemblies [4]

The structures of the full-scale test assemblies are shown in Figure 1. Timber studs are

shown in their final charred shapes (red cross-sections).

2.3

Test results

3

Insulation

3.1

Insulation types

3.1.1

Mineral wool

European standard for mineral wools EN 13162 [5] gives the properties of mineral wools,

but does not give any special properties or requirements for their protective performance

of timber structures when exposed to fire.

Depending on different protection properties for timber members in fire situations,

mineral wool products are divided into two types in EN 1995-1-2 [1] - stone wool and

glass wool.

3.1.2

Stone wool

Stone wool (rock wool) is a mineral wool manufactured predominantly from molten

naturally occurring igneous rocks [6].

Steam of air is blown through the molten rock of about 1600 °C. After formation, the

fibres are sprayed with binding agents, water repellents and mineral oil, and are passed

through an oven and are formed into the appropriate products. The final product is a mass

of fine, intertwined fibres with a typical diameter of 6 to 10 micrometers.

Traditional stone wool is not sensitive to high temperatures in a standard fire.

3.1.3

Glass wool

Glass wool is a mineral wool manufactured predominantly from sand and molten

recycled glass [6]. It consists of intertwined and flexible glass fibres, which cause it to

trap air, resulting in a low density that can be varied through compression and binder

content.

After fusion of a mixture of natural sand and recycled glass at 1450 °C, the glass that is

produced is converted into fibres. The cohesion and mechanical strength of the product is

provided by the presence of a binder that “cements” the fibres together. Ideally, a drop of

bonder is placed at each fibre intersection. This fibre mat is then heated to around 200 °C

to polymerize the resin, and is calendared to give it strength and stability.

Densities of glass wool insulations, used in structures, are usually 14 to 20 kg/m

_.

Traditional glass wool is sensitive for high temperatures in fires. When the temperature

exceeds 500 °C, there can be a fast recession of traditional glass wool insulation. This

will occur usually after the cladding failure.

3.1.4

Glass wool with high maximum service temperature

Glass wool with a high maximum service temperature is resistant to high temperatures.

This is achieved by using a patented glass compound. The material is manufactured using

similar technology to that for traditional glass wool: the difference is a higher quality of

the raw material and a higher temperature in the production process. The patented glass

compound has a very high temperature resistance.

Insulation properties of the material at normal temperatures are similar to those of

traditional glass wool. Conductivity at normal temperatures is λ=0,037 W/mK

At present, there is the only one producer of the material: Saint-Gobain Isover.

This report describes the performance of two different Ultimate glass wool products:

Ultimate UniQ Plus

(also referred to as UniQ+ ) with a nominal density of 20 kg/m

_{, and}

Ultimate UniQ,

with a nominal density of 14 kg/m

3.1.5

Tested insulations

The densities of the insulations in the tests were measured by weighing the packets in the

fire laboratory before fitting the insulation into the test structures.

Table 1 - Densities of insulation materials

Material

Densities, kg/m

Average density, kg/m

Ultimate UniQ Plus

20,5 to 21,1

21,0

Ultimate UniQ

13,8 to 14,7

14,3

Stone wool, RW4

29,0 to 30,0

29,4

Stone wool, RW3

36,2 to 37,1

36,6

Stone wool, RW2

28,9

28,9

Stone wool, RW1

27,9 to 30,5

29,4

Glass wool, KL-35

18,4

18,4

3.2

Thermal properties of insulation

The conductivity values for stone wool and glass wool is given in Figure 2.

Figure 2 - Effective conductivity of the insulations [7]

The effective values given here represent the values for insulation in panel structures at

high temperatures. Cooling effect, air penetration effect or other effects, can influence the

properties in a real structure.

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

0.5

0

200

400

600

800

1000

1200

λ

[W

/m

K

]

Temperature [C]

stone wool

glass wool

3.3

Temperature rise behind the insulation

The fire tests measured the temperature rise between the insulation and the gypsum board

on the unexposed sides. Figure 3 shows the results of the temperature measurements

behind different insulation materials, with and without protection on the fire-exposed

side.

In some tests, the insulation is protected by Type A (GtA) or Type F (GtF) gypsum

plasterboards on the exposed side. There are unprotected compositions on the

fire-exposed side when the cladding is not named in the key for Figure 3.

The temperature rise behind traditional glass wool is rapid after the cladding falls off,

although there is a noticeable difference when compared to the wall assembly with a void

cavity (in the same test): compare GtA12+void and GtA12+GW in Figure 3. Glass wool

provides a certain delay of rapid temperature rise during the recession period after the

gypsum cladding falls off.

The temperature rise behind the heat-resistant Ultimate UniQ and UniQ+ glass wools is

quite similar to stone wool products, see Figure 4.

One of the stone wool products (RW2) showed noticeably worse protection that other

stone wools, and is therefore shown with a separate curve in Figure 3. The other stone

wool products (RW1, 3, 4) performed similarly to each other, and are shown by a single

curve.

The observed difference in protection ability for different stone wool products of different

producers confirms the assumption stated before, that it may be unsafe to use existing

rules given by EN 1995-1-2 (1) for all stone wool products.

From these test results, Ultimate UniQ protects the wooden stud similarly to stone wool

RW2, which had the poorest protective performance among stone wools. Ultimate

UniQ+

showed similar protective properties to stone wool products RW1, 3 and 4.

It is obvious that gypsum plasterboard cladding provides a noticeable delay for

temperature rise. After the plasterboard falls off, temperature rise is faster for all types of

structures, and the difference in temperatures on the unexposed side of protected and

unprotected walls decreases.

Test walls with Type A gypsum plasterboard cladding have the same temperature rise

behind the insulation until failure time t

, whether insulated with Ultimate or traditional

glass wool.

The first layer on the exposed side is always the most important one in determining the

total fire resistance of a structure. It can also be seen in Figure 3 that traditional glass

wool protected by 15 mm of type F gypsum board (GtF15+GW) shows the same

temperature on the unexposed side after 60 minutes as does Ultimate UniQ+ protected by

12,5 mm of Type A gypsum board, probably fulfilling the criteria for EI60.

Figure 3 - Temperature rise measured between the insulation and gypsum board on the

unexposed side

Figure 4- Temperature rise measured behind Ultimate glass wool and stone wool

0

50

100

150

200

250

300

350

400

0

10

T

e

m

p

e

ra

tu

re

r

is

e

b

e

h

in

d

t

h

e

i

n

s

u

la

ti

o

n

[

K

]

RW 2

UniQ Plus

GtA12+void

GtA12+UniQ Plus

GtF15+air12,5+UniQ Plus(120)

Temperature rise measured between the insulation and gypsum board on the

Temperature rise measured behind Ultimate glass wool and stone wool

20

30

40

50

t

[min]

RW 1,3,4

UniQ

GtA12+GW

GtF15+GW

GtF15+air12,5+UniQ Plus(120)

Temperature rise measured between the insulation and gypsum board on the

60

3.4

Small-scale electric

heated furnace

Temperature rises behind the insulation materials differed, depending on the different

heating conditions. The same materials, produced in the same batches, were tested in the

Saint-Gobain Isover factory in Lübz

furnaces followed the standard fire exposure curve (8).

The materials tested on the electrical furnace were tested horizontally, without cladding

on the unexposed side. The materials tested on the gas

vertically, with cladding on the unexposed side.

Peaks of about 150 K on the electrical furnace show the energy increase when the heated

binder reacts. Because of the very small size of the heating chamber, without suction, the

energy from the burning binder goes directly through the product.

Figure 5 shows the remarkable difference in temperature rise behind the insulation for

two different furnaces. After one hour of heating by the gas

behind insulation (with cladding on the unexposed side) is about 100

the temperatures recorded for the corresponding tests by the electric furnace.

Figure 5- Temperature rise behind the insulation

One of the reasons for the observed differences can be that the tests using the electrical

furnace do not generate gases resulting from the burning of timber. The insulation has

scale electrical furnace vs full-scale gas

heated furnace

Temperature rises behind the insulation materials differed, depending on the different

heating conditions. The same materials, produced in the same batches, were tested in the

Gobain Isover factory in Lübz and in full-scale tests in Tallinn. Tests in both

furnaces followed the standard fire exposure curve (8).

The materials tested on the electrical furnace were tested horizontally, without cladding

on the unexposed side. The materials tested on the gas-heated furnace were tested

vertically, with cladding on the unexposed side.

K on the electrical furnace show the energy increase when the heated

binder reacts. Because of the very small size of the heating chamber, without suction, the

nergy from the burning binder goes directly through the product.

shows the remarkable difference in temperature rise behind the insulation for

. After one hour of heating by the gas-fired furnace, the temperature

behind insulation (with cladding on the unexposed side) is about 100 - 150 K higher than

the temperatures recorded for the corresponding tests by the electric furnace.

behind the insulation

One of the reasons for the observed differences can be that the tests using the electrical

furnace do not generate gases resulting from the burning of timber. The insulation has

scale gas

Temperature rises behind the insulation materials differed, depending on the different

heating conditions. The same materials, produced in the same batches, were tested in the

scale tests in Tallinn. Tests in both

The materials tested on the electrical furnace were tested horizontally, without cladding

eated furnace were tested

K on the electrical furnace show the energy increase when the heated

binder reacts. Because of the very small size of the heating chamber, without suction, the

shows the remarkable difference in temperature rise behind the insulation for

fired furnace, the temperature

K higher than

One of the reasons for the observed differences can be that the tests using the electrical

furnace do not generate gases resulting from the burning of timber. The insulation has

better properties without turbulence and convection. In real structures, this effect must be

considered.

3.5

Design models

3.5.1

Design model for floor and wall assemblies completely

filled with stone wool according to EN 1995-1-2:2004

For beams or columns protected by stone wool batts with a minimum thickness of 20 mm

and a minimum density of 26 kg/m

_{, which maintain their integrity up to 1000 °C, the}

start of charring time t

should be taken according to Equation (3.13) of EN

1995-1-2:2004 (1)

t

= 0, 07 h

(

− 20

)

ρ

(3.1)

For a thickness of 145 mm and a density of 29 kg/m

_{, the start of charring time is}

calculated as:

tch

= 47,1 min

3.5.2

Design model for Ultimate by ETH

The start of charring time for Ultimate was researched by ETH (9), and is explained as

follows (Equation [5] in [9]):

t

_ch

= 36

h

80









0,8

ρ

20









0,9

From this, the start of charring time for Ultimate UniQ Plus, with a thickness of 145 mm

and a density of 21 kg/m

_{, is:}

tch

= 60,5 min

The start of charring time for Ultimate UniQ for a thickness of 145 mm and a density of

14 kg/m

_is:

tch

= 42 min

3.5.3

Tests

Figure 6- Start time of charring behind the 145 mm insulation

One of the tested stone wools showed quite early charring behind the insulation. Charring

started 3 - 4 minutes earlier than according to the design rule in Eurocode. Temperatures

behind other tested insulations did not reach 300

The start time of charring behind the

(10 - 30 %) later than as found in

during the tests.

a) Ultimate UniQ after the test

Decrease of thickness is about 15 mm.

Figure 7 - Thicknesses of the insulations after fire tests

The thickness of the materials was me

less decrease in thickness for

of charring behind the 145 mm insulation

One of the tested stone wools showed quite early charring behind the insulation. Charring

4 minutes earlier than according to the design rule in Eurocode. Temperatures

tested insulations did not reach 300 °C during the test.

of charring behind the Ultimate UniQ insulation was 5 to 12 minutes

%) later than as found in (9). There was no charring behind Ultimate UniQ Plus

after the test

Decrease of thickness is about 15 mm.

b) RW 2 after the test

Decrease of thickness is about 20-35 mm.

Thicknesses of the insulations after fire tests

The thickness of the materials was measured after the tests, and showed 2 to 2,5 times

less decrease in thickness for Ultimate than for the reference stone wool.

One of the tested stone wools showed quite early charring behind the insulation. Charring

4 minutes earlier than according to the design rule in Eurocode. Temperatures

insulation was 5 to 12 minutes

Ultimate UniQ Plus

35 mm.

Figure 8 - Surface behind Ultimate

Figure 9 - Surface behind stone wool

The paper surface behind the insulation is slightly brown. Charring has just started.

Visually, there is no differences between surfaces protected by stone wool and by

Ultimate UniQ+.

See Figure 8 and Figure 8.

3.6

Influence of joints in the insulation

Test 2.2

Tests 2.3, 2.4

Figure 10 - Location of joints in the insulation

The behaviour of insulation at joints is different for stone wool and for Ultimate.

Presented surface pictures (Figure 13 for example) and thermo-camera pictures

(Figure 11) are taken from different sides (in the mirror).

435

1000

870

1000

1000

UniQ

UniQ+

RW4

RW2

RW1

RW3

a) Test 2.3 (ultimate), 58 min

b) Test 2.2 (stone wool), 59 min

UniQ at left, UniQ Plus at right

Figure 11 - Temperature spread on unexposed surface

The temperature on the Ultimate-insulated side is more equal over the area than is that of

the stone wool-insulated side .

a) Ultimate

_{b) Stone wool}

Figure 12 - Joints during the fire test

Figure 13 - Insulation after the 60 minutes test

Stone

wool

UniQ+

No downward movement of

remained horizontal, see Figure

temperatures.

Expansion of Ultimate at elevated temperatures was noticed during the test. Joints of

Ultimate

opened on the

fire-closed on the unexposed side.

Stone wool batts exposed directly to fire show opened joints. Batts are slightly bowing

downwards. A steel mesh was fitted to prevent fall

showed this was not needed.

3.7

Temperature profiles on the unexposed side

The temperature measurements by thermocamera during the fire test showed that the

temperature rise is more uniform for

wool.

For stone wool, the difference in temperature rise on the unexposed side of studs and

joints is greater than for Ultimate

For more temperatures recorded by thermocamera, see Annex B

3.8

Effect of air gap

Test measurements showed the cooling effect of an air gap behind cla

the temperature rise on the unexposed side: see

temperature rise behind the cladding, prolonging thermal degradation of the gypsum

board by about 3 - 5 minutes.

Figure 14 - Temperatures behind Type F gypsum boards

No downward movement of Ultimate batts was observed during the fire test. Joint lines

Figure 13. This is due to expansion of the material at high

at elevated temperatures was noticed during the test. Joints of

-exposed side on the surface, (see Figure 12), and stayed

closed on the unexposed side.

Stone wool batts exposed directly to fire show opened joints. Batts are slightly bowing

downwards. A steel mesh was fitted to prevent fall-off of stone wool batts, but the test

Temperature profiles on the unexposed side

The temperature measurements by thermocamera during the fire test showed that the

temperature rise is more uniform for Ultimate over the unexposed surface than for stone

he difference in temperature rise on the unexposed side of studs and

Ultimate

.

For more temperatures recorded by thermocamera, see Annex B

Effect of air gap

Test measurements showed the cooling effect of an air gap behind cladding in reducing

the temperature rise on the unexposed side: see Figure 14. The air gap avoids the local

temperature rise behind the cladding, prolonging thermal degradation of the gypsum

utes.

Temperatures behind Type F gypsum boards

batts was observed during the fire test. Joint lines

. This is due to expansion of the material at high

at elevated temperatures was noticed during the test. Joints of

), and stayed

Stone wool batts exposed directly to fire show opened joints. Batts are slightly bowing

off of stone wool batts, but the test

The temperature measurements by thermocamera during the fire test showed that the

over the unexposed surface than for stone

he difference in temperature rise on the unexposed side of studs and

dding in reducing

. The air gap avoids the local

temperature rise behind the cladding, prolonging thermal degradation of the gypsum

Figure 15- Temperature measured on the unexposed side

GtF 15 GtA 13

GW 145 Ultimate 145

GtF 15 GtA 13

Figure 16- Temperatures on unexposed side for different structures

4

Analysis of charred cross

4.1

Test walls

4.1.1

General

Most of the test walls were cons

to eliminate the uncertainty of failure time of cladding.

The EN 1995-1-2 [1] rules for insulated structures were developed based on the

assumption that all stone wool products give the same resul

materials' densities. Since, in the past, products may have been optimised for other

aspects (e.g. production costs etc.), and no standard for verification of fire resistance in

fire resistance tests for mineral wool exists, it ca

wools are equal.

Stone wool was used to compare results with the existing model in

stone wool products was the same product as the reference, but produced in 2009.

Temperature measured on the unexposed side

GW 145 Ultimate 145

GtF15

Air 12,5

Ultimate 120

Air 12,5

GtF15

Temperatures on unexposed side for different structures

Analysis of charred cross-sections

Most of the test walls were constructed without cladding on the fire-exposed side in order

to eliminate the uncertainty of failure time of cladding.

rules for insulated structures were developed based on the

assumption that all stone wool products give the same results, depending only on the

materials' densities. Since, in the past, products may have been optimised for other

aspects (e.g. production costs etc.), and no standard for verification of fire resistance in

fire resistance tests for mineral wool exists, it can no longer be assumed that all stone

Stone wool was used to compare results with the existing model in [10]. One of the tested

stone wool products was the same product as the reference, but produced in 2009.

exposed side in order

rules for insulated structures were developed based on the

ts, depending only on the

materials' densities. Since, in the past, products may have been optimised for other

aspects (e.g. production costs etc.), and no standard for verification of fire resistance in

n no longer be assumed that all stone

. One of the tested

stone wool products was the same product as the reference, but produced in 2009.

All studs were cut from the same batch of raw material. Studs were strength graded, Class

C24, produced by Stora Enso Imavere sawmill in Estonia. Characteristic density of the

studs is 450 to 500 kg/m

_{, and their moisture content was 10 to 12 % before the test.}

4.1.2

Determination of properties of charred cross-sections

Residual cross-sections of the studs were measured by an ATOS II optical scanner [17].

Section properties (area, section modulus, charring depth etc.) were determined by

AutoCAD.

Detailed section properties of charred cross-sections are given in Annex A

Figure 17 - Definition of stud numbers of tested walls. View from the fire side.

2 9 6 0 2930 2 8 7 0 265 600 600 600 600 265 555 198

1

2

3

4

5

6

7

2.1.

3

a) Charring from one side

Figure 18 - Description of symbols for charred cross

4.2

Charring rate

4.2.1

Design in accordance with EN 1995

Rules for fire design of wall and floor assemblies insulated by stone wool and glass wool

are given in Annex C of EN 1995

Figure 19- Charring of timber studs with and without protection

Charring rate is counted as

β

= k

k

k

β

where

β

= 0, 65

mm

min

one-dimensional charring rate

kp

– protection factor

ks

– cross-section factor

kn

– factor to convert actual cross

z

1

z

2

d

c

h

a

r

d

c

h

a

r,

s

2

d

c

h

a

r,

s

1

ds2,max

b) Charring from three sides

on of symbols for charred cross-sections

Charring rate

Design in accordance with EN 1995-1-2:2004

Rules for fire design of wall and floor assemblies insulated by stone wool and glass wool

EN 1995-1-2 [1].

1- Unprotected membe

2,3 - Initially protected

members.

2 - Charring starts at t

reduced rate when protection

is still in place

3a After protection has fallen

off, charring continues at

increased rate

3b Char layer acts as a

protection and charring rate

decreases

Charring of timber studs with and without protection

dimensional charring rate

ctual cross-section into notional rectangular cross-section

z

1

z

2

d

c

h

a

r

ds1,40

ds2,40

ds1,max

ds2,max

4

0

Rules for fire design of wall and floor assemblies insulated by stone wool and glass wool

Unprotected members

Initially protected

Charring starts at t

at a

reduced rate when protection

3a After protection has fallen

off, charring continues at

3b Char layer acts as a

protection and charring rate

Protection factor k

takes the protection phase into account. When cladding is still in

place, factor k

is valid. When cladding has fallen off, factor k

applies for the

post-protection phase.

According to [1], glass wool-insulated wall and floor assemblies are assumed to lose their

load-carrying capacity when the cladding on the fire-exposed side has fallen off.

For stone wool-insulated assemblies and assemblies with void cavities, design rules are

given for the post-protection phase in [1].

Protected charring phase

t

≤

≤

≤t≤

≤

≤

≤tf

Before failure of the protective cladding, there is no difference in the fire behaviour of

assemblies with stone wool or with glass wool.

Phase 2 charring rates of the timber member given in Table 3.1 of EN 1995-1-2 [1]

should be multiplied by an insulation factor k

.

Post-protection charring phase

t > t

Provided that the cavity insulation is made of stone wool batts and remains in place after

failure of the cladding, the post-protection factor k

should be calculated as (Subclause

C.2.1 (5) at (1)):

According to (1), the model is a reasonable approximation of charring depths up to

40 mm.

Where the cavity insulation is made of glass wool, failure of the member should be

assumed to take place at time t

, (Subclause C.2.1 [6] in [1]), which is a conservative

approach. Since only a few tests were done in the past, no model data was available for

inclusion in EN 1995-1-2:2004 in the existing version.

Protection factors k

and k

have not not been verified in this study.

4.2.2

Cross-section factor

k

s

Cross-section factor k

takes the width of the cross-section into account. Charring is faster

when the cross-section is smaller, due to two-dimensional heat flux within the member.

Size factor k

is determined by charring for 30 mm charring depth from the narrow side,

as is done in [10].

ks

=

β

/

β

Thus, one-dimensional charring rate for 30 mm charring is:

β

= 30 / t

where

tch,30

– time until charring of 30 mm occurs

3

0, 036

1

k

₌

t

₊

(4.4)

(4.3)

(4.2)

β

= 0,65 mm/min

The following includes the charred cross-sections from Tests 2.2, 2.3 and 2.4. All of them

were tested without cladding on the fire-exposed side.

EN 1995-1-2 [1] gives a value of k

= 1,3 for a cross-section width of 45 mm.

Figure 20 compares the results with König [10].

Figure 20- Charring depth versus time. Comparison of present results with König [10]

Figure 21- Comparison of trendlines of old and new results of charring depth of wall

studs insulated by stone wool.

d

= 0,0091 t

_{+ 0,549 t}

0

10

20

30

40

50

60

70

0

10

20

30

40

50

60

d

c

h

a

r

[m

m

]

t

[min]

Tests S1, S3-S7

UniQ+

UniQ

RW1

RW2

RW3

RW4

UniQ+ (GtA12,5)

45 x 145

d

= 0,0091 t

_{+ 0,55 t}

d

= 0,0029t

+ 0,92t

0

10

20

30

40

50

60

70

0

10

20

30

40

50

60

d

c

h

a

r

[m

m

]

t

[min]

RW tests (König)

RW tests (Just)

45 x 145

Figure 22- Charring depth for studs protected by stone wool and Ultimate respectively

Figure 23- Charring rates for studs insulated by Ultimate and stone wool.

Polynomial trend lines of test results in Figure 21 with reference material – stone wool –

show the difference between previous and present results:

The charring rate is higher; the reason for that could probably be a change in the

composition of the stone wool by the producers during the past.

Results from [10] lead to the following relation:

dchar

= 0,0091t

+0,55t

The present research found the charring depth for 45 mm wide cross-section, protected

with stone wool as follows:

dchar

= 0,0029t

+0,92t

Charring of studs protected by Ultimate is described as

UniQ Plus

dchar

= 0,006t

+0,77t

d

= 0,006t

_{+ 0,771t}

d

= 0,0052t

_{+ 0,911t}

d

= 0,0029t

_{+ 0,917t}

0

10

20

30

40

50

60

70

0

10

20

30

40

50

60

70

80

d

c

h

a

r

[m

m

]

t

[min]

ks=1,6

UniQ+

UniQ

RW

Poly. (UniQ+)

Poly. (UniQ)

Poly. (RW)

45 × 145

45 x 145

d

= 1,074t

d

= 1,168t

d

= 1,067t

0

10

20

30

40

50

60

70

0

10

20

30

40

50

60

70

80

90

d

c

h

a

r

[m

m

]

t

[minutes]

ks=1,6

UniQ+

UniQ

RW

Linear (UniQ+)

Linear (UniQ)

Linear (RW)

45 × 145

45 x 145

(4.7)

(4.6)

(4.5)

UniQ

dchar

= 0,0052t

+0,91t

For simplification, a linear approach for charring depth has been used. Linear trendlines

of stone wool and Ultimate products are shown in Figure 23.

Ultimate UniQ Plus

provides the same protection as stone wool. Ultimate UniQ gives the

same protection until 40 mm charring depth, after which charring is slightly faster.

A charring depth of 40 mm is taken as a basis for design requirements. A greater charring

depth for relatively small wall or floor studs would be irrelevant, because of loss of

bearing capacity.

Figure 24- Charring depth versus time. Test results and recommended value for k

=1,6.

The charring rate of structures without cladding is calculated as

β

=k

β

Charring depth is calculated as

dchar

=

β

t

In the present standard, EN 1995-1-2 [1], the value for k

is 1,3.

The charring depth for 45 mm wide wall studs protected with stone wool is expressed as:

dchar

= 0,845 t

According to the test results, the proposal for charring depth of a 45 mm wide stud when

protected by Ultimate UniQ or UniQ Plus should be

dchar

= 1,04 t

The same charring depth is proposed for use for the same studs insulated by stone wool.

0

10

20

30

40

50

60

70

0

10

20

30

40

50

60

70

80

90

d

c

h

a

r

[m

m

]

t

[min]

ks=1,3

ks=1,6

UniQ+

UniQ

RW

45 × 145

45 x 145

d

=1,04 t

d

=0,845 t

(4.12)

(4.11)

(4.10)

(4.9)

(4.8)

The recommended new value for cross

ks

= 1,6.

Figure 25 – Coefficient k

In [10], it is given as

Expression (4.15) assumes a linear relationship between

conservative for d

< 30 mm and non

the load resistance is normally exhausted).

New results from [2] show that stone wool from various producers provides somewhat

less fire protection of the wide sides of the timber member than reported in

Expression (4.15) should therefore be replaced with

This expression is illustrated in

The expression can be used for Ultimate glass wool with a minimum density of 14

and for stone wool with a minimum density of 26 kg/m

4.2.3

Conversion factor

Conversion factor k

converts the irregular charred cross

rectangular cross-section.

0,000167

0,029

2,27

for 38 mm

90 mm

1

for

90 mm

b

b

b

k

= 



−

+

≤ ≤



0,00023

0,045

3,19

for 38 mm

90 mm

1

for

90 mm



₋

₊

_{≤ ≤}

= 



b

b

b

k

The recommended new value for cross-section factor for 45 mm wide cross-sections is

Expression (4.15) assumes a linear relationship between d

and time, which is slightly

mm and non-conservative for d

> 30 mm (for d

> 30

ormally exhausted).

show that stone wool from various producers provides somewhat

less fire protection of the wide sides of the timber member than reported in [10].

Expression (4.15) should therefore be replaced with

ion is illustrated in Figure 25.

The expression can be used for Ultimate glass wool with a minimum density of 14

and for stone wool with a minimum density of 26 kg/m

_.

Conversion factor

k

n

converts the irregular charred cross-section into a notional

0,000167

0,029

2,27

for 38 mm

90 mm

1

for

90 mm

b

b

b

b

−

+

≤ ≤

>

0,00023

0,045

3,19

for 38 mm

90 mm

1

for

90 mm

−

+

≤ ≤

>

b

b

b

b

sections is

and time, which is slightly

30 mm

show that stone wool from various producers provides somewhat

The expression can be used for Ultimate glass wool with a minimum density of 14 kg/m

(4.14)

(4.13)

Key:

1

Solid timber member (stud or joist)

2

Cladding

3

Insulation

4

Residual cross-section (real shape)

5

Char layer (real shape)

6

Equivalent residual cross-section

7

Char layer with notional charring depth

Figure 26– Charring of timber frame member (stud or joist): a. Section through

assembly. b. Real residual cross-section and char layer. c. Notional charring depth and

equivalent residual cross-section

Determining conversion factor k

The properties of the residual cross-section: see Annex A.

Section modulus of residual cross-section is counted as

Wfi

= min (I

/z

)

Moments of inertia and distances to centroid are determined by measuring of residual

cross-sections. See tables in Annex A.

Charring on the narrow side is counted as

dchar

=h-h

Corresponding rectangular cross-section is calculated from the equation

Wrec

=W

b

=45 mm,

h

=

6W

b

Wfi

– actual residual cross-section

Wrec

– notional rectangular cross-section

hrec

– height of notional rectangular cross-section

Notional charring

dchar,n

=h-h

b

d

h

6

7

b

5

4

a)

b)

c)

2

d

(4.18)

(4.17)

(4.16)

(4.15)

Conversion factor is the following relation:

kn

=d

/d

Afi

– notional rectangular cross-section corresponding to equal values of section modulus

of the real and notional rectangular sections.

Cross-section properties are shown in Annex A.

The k

value changes during char development. EN 1995-1-2 gives a simplification,

kn

=1,5, which gives overestimated results in the early phase of a fire and more

conservative results with thicker charring depth.

Relations from the present test results match well with J.König’s research [10]. See

Figure 27, where the present results are added to the results from [10].

There is a good correlation between the present results and results from [10].

Figure 27- Reduction of section modulus versus charring depth ratio.

0.0

0.2

0.4

0.6

0.8

1.0

0.0

0.2

0.4

0.6

0.8

W

/W

d

/h

GtF

GtF+GtA

Unprotected

2xGtF

1,6

1,4

1,3

Test (König)

S1, S3

RW tests (Just)

UniQ tests (Just)

UniQ+ tests (Just)

45 x 145

Figure 28- Reduction of section modulus versus charring depth ratio. Test results

Figure 29- Reduction of moment of inertia versus charring depth ratio. Test results

0.0

0.2

0.4

0.6

0.8

1.0

0.0

0.2

0.4

0.6

0.8

W

/W

d

/h

1,5

Test RW

Test UniQ

Test UniQ+

45 ξ 145

0.0

0.2

0.4

0.6

0.8

1.0

0.0

0.2

0.4

0.6

0.8

I

/I

d

/h

1,5

Test RW

Test UniQ

Test UniQ+

45 x 145

Figure 30 - Reduction of cross-section area versus charring depth ratio. Test results

The cross section insulated with heat-resistant glass wool show similar relations with the

stone wool-insulated cross section when converting an irregular residual cross section

into a notional rectangular cross-section.

Relations for section modulus, moment of inertia and cross-section area are shown in

Figure 28 to Figure 30. Each point represents a different section.

4.2.4

Online determination of charred cross-sections

The development of residual cross-section was measured in the presented test series by

thermocouples on the sides of, and inside, the cross section. The following describes the

formation of each measured cross section.

Residual cross sections are calculated by drawing polynomials through three points

(thermocouple locations) and adding the rectangular unburnt part of the cross section.

Values between thermocouples are interpolated when necessary.

See Figure 31.

0.0

0.2

0.4

0.6

0.8

1.0

0.0

0.2

0.4

0.6

0.8

A

/A

d

/h

1,5

Test RW

Test UniQ

Test UniQ+

45 x 145

Figure 31- Principle of calculating cross-section properties during a fire test

Figure 32- Change of cross-section area versus charring depth ratio during the test.

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0

0.1

0.2

0.3

0.4

0.5

0.6

A

/A

d

/h

A

_fi

/A

uniq+_2.3

UniQ_2.3

uniq+_2.4

RW4_2.4

RW3_2.2

RW1_2.2

RW4_2.2

RW2_2.2

kn=1,5

k

_n

=1,5

Figure 33- Change of section modulus versus charring depth ratio during the test.

The online measurements confirm the theory that factor k

is on the unsafe side for small

charring depths, and on the safe side for greater charring depths.

Taking into account the safety factors and load cases, the important area in fire safety

design is W

/W = 0,2 to 0,4. For that zone, a value of k

= 1,5 is a good approximation.

4.3

Charring on the sides of cross-sections

Figure 34 – Relationship between charring on the sides and in the middle.

Relations between charring on the sides and charring depth in the middle of the narrow

side of a cross-section is shown in Figure 34.

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0

0.1

0.2

0.3

0.4

0.5

0.6

W

/W

d

/h

W

_fi

/W

uniq+_2.3

UniQ_2.3

uniq+_2.4

RW4_2.4

RW3_2.2

RW1_2.2

RW4_2.2

RW2_2.2

'kn=1,5

k

n

=1,5

1.00

1.50

2.00

2.50

3.00

3.50

0

10

20

30

40

50

60

d

/d

t

[min]

2.3_3 UniQ +

2.4_3 UniQ+

2.3_5 UniQ

2.4_5 RW 4

2.2_5 RW4

2.2_3 RW3

develops more slowly when protected by Ultimate UniQ+ on the wide faces of the

cross-section when compared to the performance of Ultimate UniQ and stone wool.

The difference plays a role only at the early stage of charring. The difference between

relations d

/d

for Ultimate and stone wool has disappeared after about 60 minutes

5

Comparison with other test results

Figure 35 – Start of charring times compared to safe design equation at

1995-1-2 [1]

Figure 35 is a comparison of the start of charring times as given in EN 1995

SP report [11]. Charring times from the test results presented here show that the rules of

EN 1995-1-2:2004 are too optimistic. On the other hand, the start of charring times in the

present tests are bigger from conservative rules

SP report [11].

Figure 36 shows a comparison of failure times of gypsum cla

and 2.6 with the times reported in SP report

Figure 36 –Failure times compared to safe design equations in

Charring rates of the tested woods are ca 20

design rules. Figure 37 shows the actual charring as well as EN 1995

size factor of k

= 1,3 and a proposal with size factor of

Comparison with other test results

rring times compared to safe design equation at [11]

is a comparison of the start of charring times as given in EN 1995-1-2

. Charring times from the test results presented here show that the rules of

2:2004 are too optimistic. On the other hand, the start of charring times in the

bigger from conservative rules according to analysis of the data base in

shows a comparison of failure times of gypsum claddings from tests 2.1, 2.5

and 2.6 with the times reported in SP report [11].

Failure times compared to safe design equations in [11]

Charring rates of the tested woods are ca 20 - 30% greater than the EN 1995

shows the actual charring as well as EN 1995-1-2 results with a

1,3 and a proposal with size factor of k

= 1,6.

and EN

2 [1] and

. Charring times from the test results presented here show that the rules of

2:2004 are too optimistic. On the other hand, the start of charring times in the

according to analysis of the data base in

ddings from tests 2.1, 2.5

30% greater than the EN 1995-1-2 [1]

2 results with a

Figure 37 - Charring rate of tested wall studs

6

Proposals for fire design

6.1

Floor joists and wall studs in assemblies of which

the cavities are completely filled with

Ultimate

glass wool or with stone wool

The design model described in EN 1995-1-2:2004 Annex

C

for timber frame

assemblies

insulated with stone wool is applied also for

timber frame assemblies

insulated with Ultimate glass wool, except in respect of the requirements for densities.

6.1.1

Charring rate

Charring rate is counted as

β

= k

k

k

β

(6.1)

where

β

= 0, 65

mm

min

one-dimensional charring rate

kp

– protection factor

ks

– cross-section factor

kn

– factor to convert actual cross section into a notional rectangular cross section

Protected charring phase

t

≤

≤

≤t≤

≤

≤

≤tf

For walls and floors filled with Ultimate insulation, traditional glass wool or stone wool,

the protection factor k

should be taken as:

0

10

20

30

40

50

0

10

20

30

40

50

60

d

[m

m

]

time [min]

Uni Q

UniQ+

RW

EC5-1-2

b=0,975 (ks=1,5)

b=1,04 (ks=1,6)

EN1995-1-2

β =0,845 mm/min

k

=1,3

k

=1

β

=0,65 mm/min

β =1,04 mm/min

k

=1,6

k

=1

β

=0,65 mm/min