You are on page 1of 8

Experimental and Numerical Investigations of Fire

Resistance of Novel Timber-Concrete-Composite Decks


R. Meena 1; M. Schollmayer 2; and T. Tannert 3

Abstract: Timber-concrete-composite (TCC) systems, commonly with timber at the bottom and concrete at the top, are efficient solutions
for floors because of their advantages compared to pure timber or pure concrete slabs, one of the advantages being better fire resistance. This
Downloaded from ascelibrary.org by University of Waterloo on 09/03/21. Copyright ASCE. For personal use only; all rights reserved.

paper presents experimental and numerical investigations on the fire resistance of a novel TCC setup with timber beams at the top and a
concrete layer at the bottom. This setup has several advantages in facilitating the connection to the surrounding structure and mechanical
installations. The results from fire tests on a subsection of the system are reported. The concrete side of one test specimen was subjected to a
fire according to the standard temperature-time curve and proofed to achieve a 60-min fire rating. Subsequently the temperature behavior was
numerically modeled by performing a transient thermal analysis with temperature dependent material properties. The temperature profiles for
the TCC system was accurately modeled allowing for the validated model to be used for future optimizations and predicting the fire resistance
ratings of optimized systems. DOI: 10.1061/(ASCE)CF.1943-5509.0000539. © 2014 American Society of Civil Engineers.

Introduction action and the timber must be strong enough to resist tensile
stresses due to bending. The location of the neutral axis depends
Timber Concrete Composites purely on the geometry of the system and material properties of the
timber with respect to the concrete. Since the strength of timber in
Hybrid systems that integrate different materials can significantly tension is well-documented, the connector stiffness has been the
increase the applications of timber in structures beyond current focus of recent research. Several means of connecting timber to
limitations regarding maximum number of stories. As an example, concrete are widely documented. The connectors can be catego-
timber-steel hybrid systems combine the high strength and ductility rized as (1) mechanical fasteners such as self-tapping screws, bolts,
of steel with the high stiffness, high compressive load capacity, or the proprietary HSK system (Clouston et al. 2005); (2) connec-
low weight, and low environmental footprint of timber. Timber- tions based on mechanical interlocking (notches); and (3) most
concrete-composite (TCC) systems were primarily developed recently, adhesive connections (Brunner et al. 2006; Negrão et al.
to refurbish historical buildings in Europe (Yeoh et al. 2011). 2010). Glued-in steel connectors, such as the HSK system, have
Structural and nonstructural benefits over generic timber floors, proven to be the best option in terms of the system’s overall
such as increased floor capacity and stiffness, shallower floor depth strength and stiffness (Yeoh et al. 2011).
and reduction in floor weight, and improved acoustic and fire per- Research on many aspects and components of TCC systems
formance have led to increased attention and widened application [varying connection types, influences of timber and concrete
of TCC systems. Additional advantages when compared to a con- material properties, short and especially long-term behavior
crete slab are found in the relatively rapid erection of the system (e.g., Fragiacomo 2005, 2006), and vibration behavior] while
due to the use of the timber element as formwork, more economical well-documented is also ongoing. Different design methods have
gravity systems due to TCC’s lighter weight, lower embodied been proposed for TCC floors; the simplest method [called the
energy, and reduced carbon dioxide emissions (Clouston and γ method and referred to in EN 1995-1-1:2004 (CEN 2004)] is
Schreyer 2008). based on the effective bending stiffness accounting for composite
In order to take full advantage of the benefits of TCC, the most action at the interface interlayer (Yeoh et al. 2011).
efficient TCC deck has its neutral axis at the interface between the
two components such that all of the concrete is in compression and Fire Resistance of Timber Concrete Composites
all of the timber is in tension (Yeoh et al. 2011). The connection
between the two materials must be stiff enough to ensure composite Similar to regulations in many other countries, the National
Building Code (NBC) of Canada (NRC 2010) prescribes the re-
1
Research Assistant, Timber and Composite Construction, Bern Univ. quirements for fire rating and fire safety of structures. In NBC,
of Applied Sciences, Solothurnstrasse 102, Biel 2500, Switzerland. E-mail: timber falls under combustible construction, whereas concrete is
iitbrahul@gmail.com noncombustible. In order to ensure that a TCC system satisfies
2
Professor, Timber and Composite Construction, Bern Univ. of Applied the requirements of NBC, the clauses for both materials must be
Sciences, Solothurnstrasse 102, Biel 2500, Switzerland. E-mail: martin considered to ensure code compliance. For application in Canada,
.schollmayer@bfh.ch it must be demonstrated that a TCC system can provide adequate
3
Assistant Professor, Dept. of Wood Science/Dept. of Civil Engineer- fire separation as well as maintain its structural integrity for the
ing, Univ. of British Columbia, 2424 Main Mall, Vancouver, BC, Canada
required amount of time (60-min fire separation for most occu-
V6T1Z4 (corresponding author). E-mail: thomas.tannert@ubc.ca
Note. This manuscript was submitted on April 30, 2013; approved on pancy categories).
November 11, 2013; published online on November 13, 2013. Discussion To build fire resistant structures is one of the imperative tasks in
period open until December 18, 2014; separate discussions must be submitted structural design. The idea of protecting structures against the
for individual papers. This paper is part of the Journal of Performance of effects of fire evolved late in the nineteenth century in order to re-
Constructed Facilities, © ASCE, ISSN 0887-3828/A4014009(8)/$25.00. duce losses of human life and property (Gosselin and Lie 1987).

© ASCE A4014009-1 J. Perform. Constr. Facil.

J. Perform. Constr. Facil., 2014, 28(6): A4014009


With this intention, designers started to make greater use of fire (O’Neill 2009). Current design models, utilizing the γ-method, take
resistive materials to minimize both the spread of fire and the into account the loss in cross section due to charring of wood and
amount of damage sustained by structures. The use of timber the temperature-dependent reduction of strength and stiffness of the
and its wood derivatives was limited to low-rise buildings. In recent uncharred section (Frangi et al. 2008). With a reduced cross section
years, significant research has been undertaken on the fire resis- to account for the reduction in strength and stiffness of timber in
tance of wood (e.g., Babrauskas 2002; Frangi et al. 2009), concrete heat-affected zones, a designer may use cold strength and stiffness
(e.g., Kodur et al. 2004; Zadrazil et al. 2004), and TCC structures properties of timber. Although fire resistance time charts, based on
(e.g., Frangi et al. 2010; Branthwaite 2010). These experimental charring rates of double laminated veneer lumber (LVL) panel con-
and numerical investigations were carried out on individual figurations, are reported in the literature, the need for more tests to
elements as well as full-scale structures and allow for alternate be performed to improve the reliability of the tables is recognized
solutions using TCC systems. (O’Neill 2009).
By utilizing the relevant qualities of both materials, TCC floors Different fire exposures can be used to test structural elements;
have been documented to perform better in fire than either concrete most commonly applied is the ISO 834-1 (ISO 1999) standard-
or timber floors; the fire resistance of TCC floors can be 90 min or temperature-time-curve (STtC). The STtC represents a fully devel-
Downloaded from ascelibrary.org by University of Waterloo on 09/03/21. Copyright ASCE. For personal use only; all rights reserved.

more (Natterer 2002). The performance of concrete in fire has been oped fire in a compartment. Although widely used in fire tests to
documented through a state of the art report (Fletcher et al. 2007). rate structural elements, the STtC bears little resemblance to a real
Concrete owes its desirable long-term fire performance to its low fire temperature-time history and the available research for conven-
thermal diffusivity. Reinforced concrete, however, quickly loses tional construction assemblies has been a mixed contribution for
strength through spalling (scaling damage along its surface) and the design profession. On one hand, it has led to significant savings
may even be explosive for high-strength concrete. in construction by reducing the need for costly fire-resistance tests;
The performance of timber in fire, in contrast, is very predict- on the other hand, however, it has also severely restricted design
able and indeed desirable. Unlike steel, which heats up and loses innovations by enticing engineers to specify already tested designs
structural strength throughout its entire cross section, timber chars. instead of designing new systems. Other test standards, e.g., the
When timber is heated externally, its inherent moisture tends to temperature curve according to ASTM E119 (ASTM 2003) are
evaporate between 100 and 120°C. Above 200°C, timber undergoes more severe than STtC but longer duration tests seem to deliver
a pyrolysis reaction (thermochemical decomposition of organic only negligible differences in results (Harmathy et al. 1987).
material at elevated temperatures without the participation of oxy- Additional protection methods to increase the safety of TCC
gen), generating more combustible products. After these products systems in fire include the use of gypsum boards and sprinklers.
have burned, a layer of char is left on the surface that protects the Gypsum boards absorb heat due to their high moisture content
remaining wood from external heat exposure and penetration of and as a consequence they slow down heat transfer (Peng 2010).
oxygen (Peng 2010). While this charring layer slowly creeps into The use of sprinklers on fire resistance does not need explanation;
the remaining cross section, the charred layer, due to its low thermal nonetheless NBC permits the lessening of fire resistance time
conductivity, also protects the remaining section, allowing it to re- requirements if sprinklers are installed in a building.
tain all structural properties. The charred layer has neither strength
nor rigidity, and only the layers underneath continue to carry load Objective
(White 2000), but the char layer in TCC systems provides insula-
Given the state of the art, it may be concluded that research data on
tion to protect the concrete and the connectors against high temper-
the fire resistance of TCC systems is still in its infancy (Yeoh et al.
atures (Frangi et al. 2008).
2011). Furthermore, all of the research performed on TCC decks
To establish the capacity of a TCC member during and after a
was on systems with wood in tension and concrete in compression,
fire, the charring rate of timber at the connection must be known.
and with the wooden beams exposed to fire.
Charring rates from 0.55–0.72 mm=min are reported in the litera-
This paper presents experimental and numerical investigations
ture (White 2000; Bailey 2005; Lane 2005; O’Neill 2009; Tsai et al.
on the fire resistance of a novel TCC deck. The primary objective is
2013). Factors that affect the charring rate are the exposure to fire
to establish the fire resistance rating of the TCC system based on a
(direct or indirect), the connection to the concrete, and timber
fire test according to the ISO 834-1 (ISO 1999) STtC. The secon-
element geometry (the charring rate is significantly higher for cor-
dary objective, which is contingent on the primary objective, is to
ners that are exposed to fire from two sides). While long and narrow
validate a numerical model to provide a basis for parameter opti-
timber elements are more efficient in reducing tension forces in
mizations and predicting fire resistance ratings of optimized
TCC, such a geometry is less desirable for fire resistance because
systems.
larger built up sections have a longer fire rating (O’Neill 2009).
Research on the charring rates for double laminated beams indi-
cated that glued connections had the least amount of section loss Experimental Investigation
although screwed connections achieved a similar performance
(Tsai et al. 2013). When metal connectors are exposed, fire resis-
tance is significantly lower (Peng 2010). The concrete layer is non- Novel Timber-Concrete-Composite System
combustible and tends to expand under increased temperature. This A novel TCC system was developed in the framework of the project
expansion is beneficial, acting as a protective cladding at the TCC Swiss wood-concrete deck at the Bern University of Applied
interface, reducing the effect of temperature and delaying the start Sciences, Switzerland. The goal was to design and test a completely
of charring (Yeoh et al. 2011; O’Neill 2009). prefabricated floor system that integrates all structural and mechan-
Previous research successfully combined analytical structural ical functions, e.g., floor heating and cooling, and sprinkler system.
models and three-dimensional (3D) heat transfer numerical models To overcome the commonly encountered problem of connecting a
to predict the behavior and resistance of timber connections in TCC floor to a timber wall, an unconventional setup with timber
fire (Peng 2010). Spreadsheets to estimate fire resistance of TCC beams on top and a concrete layer on the bottom was developed.
floor systems were provided based on the fact that the capacity of The system (Fig. 1) consists of 6 × 3 m elements, which can be
a TCC in a fire depends on the tensile capacity of the timber delivered to site ready for installation.

© ASCE A4014009-2 J. Perform. Constr. Facil.

J. Perform. Constr. Facil., 2014, 28(6): A4014009


concrete that enables rapid and economic production is used.
The structural design of the presented novel TCC system is ad-
dressed in a companion paper (Hehl et al. 2013). Thermocouple
wires that can activate the system, to regulate the building temper-
ature, are fixed to the individual elements of the TCC system.
One specific advantage of the system is the fire protection on the
lower side of the floor assembly by means of the concrete layer.
There are, however, critical aspects of the system, as follows:
(1) the thin concrete coverage of the reinforcement which when
exposed to high temperature during a fire results in high risk of
spalling, (2) the increased risk of concrete chipping due to the
use of high-strength concrete, (3) the exposure of the thermocouple
wires to high temperature due to the proximity to the lower surface,
(4) the risk of losing the mechanical connection between the timber
Downloaded from ascelibrary.org by University of Waterloo on 09/03/21. Copyright ASCE. For personal use only; all rights reserved.

and the concrete during the fire, and (5) the risk that the timber
Fig. 1. Schematic of the novel TCC system with notch detail
ignites and causes an open fire in the cavity.

This novel setup, though at first it seems counterintuitive, pro- Test Specimen Description
vides most of the advantages of traditional TCC floors; specifically The primary objective of the research reported in this paper was to
it is lighter weight compared to a concrete floor and has better establish the 60-min fire resistance rating (REI 60) of the TCC
acoustical properties than a lightweight timber floor. Additionally, system based on a fire test according to the ISO 834-1 (ISO 1999)
the system significantly improves (1) the possibilities to integrate STtC. The specimen fabricated for the fire tests was a full-scale
mechanical installations (using the concrete layer as a thermoactive subsection (940 × 740 mm) of the floor system and consisted of
element), (2) the connection to timber structural systems, and the reinforced concrete slab connected to two timber beams (Fig. 2).
(3) the fire resistance of the system. The depth of the beam penetration into the slab was 30 mm and
In a statically determinate single-span application, the timber is each beam was connected to the slab with eight notches. Table 1
in compression, where it has similar strength as in tension, but lists the test specimen specifications.
favorably behaves in a ductile manner and not a brittle manner. The To produce the test specimen, the timber beams were cut to di-
concrete is in tension, which means that a reinforcement mesh is mension from Glulam GL24h, then grooves were cut in the wooden
required to carry the tension force. The concrete is held in place by beams. The bending strength f m;k and stiffness E0;mean according to
means of grooves milled into the timber beams; the shear connec- SIA 265 (SIA 2004) can be assumed to be 24 and 11,600 N=mm2 ,
tion between timber and concrete is achieved by mechanical inter- respectively. The reinforcement bars and the multilayer pipes were
locking through grooves (Fig. 1). The system is completed with a fixed into the mould (Fig. 3) and then the beam elements were
LVL panel on top which serves as base plate for the floor built up. placed into the mould. The multilayer pipe was a composite plastic/
The 80-mm concrete layer in the tension zone is relatively thin metal pipe for surface heating, radiator connection, and drink-
and therefore must be reinforced. Self-compacting high-strength ing water, and was specified according to the manufacturer’s

Fig. 2. Transverse cross section of the test specimen (all dimensions in millimeters)

© ASCE A4014009-3 J. Perform. Constr. Facil.

J. Perform. Constr. Facil., 2014, 28(6): A4014009


Table 1. Specifications of Fire Test Specimen specifications (HakaGerodur 2011) to withstand a maximum oper-
Element/material Dimensions/numbers ating temperature of 95°C and a maximum operating pressure of
1 MPa. One of the pipes was filled with water and closed, while
Specimen 0.94 m L × 0.74 m W × 0.21 m H a
Timber beam GL24h, BSH 160/80–glulam
the other pipe was kept open. All four sides of the specimen were
Concrete slab Self-consolidating concrete, covered with rock wool for insulation. Thermocouple wires were
C60/65 attached at 27 locations (Table 2) of the specimen to measure the
Numbers of steel bars Nine lateral × seven temperature of the involved materials at different times during the
longitudinal fire test. Precast self-consolidating concrete (C60/65) was poured
Steel bar diameter (mm) 5 into the mould and then the specimen was stored for 51 days until
Spacing between steel bars (mm) 100 testing.
Concrete cover for reinforcement (mm) 25
Multilayer metal pipe diameter, 16
outer (mm) Test Methods
Multilayer metal pipe thickness (mm) 2
Downloaded from ascelibrary.org by University of Waterloo on 09/03/21. Copyright ASCE. For personal use only; all rights reserved.

a Experimental investigations were conducted on the TCC system at


H = height; L = length; and W = width.
the fire test facility of the Swiss Federal Laboratories for Materials
Science and Technology (EMPA) in Dübendorf, Switzerland, in
collaboration with ETH Zürich. The test specimen [Fig. 4(a)] was
fixed into an oven (1,000 × 800 mm) and tested for 60 min under
the ISO 834-1 (ISO 1999) STtC condition.
According to standard procedure, only the bottom side of the
element was exposed to the fire. Temperature profiles were ob-
tained from the thermocouples, which were fixed to the specimen
at different locations and connected to a temperature-measuring

Fig. 3. Reinforcement and pipes fixed into the mould (open pipe on
the left)

Table 2. Location of Thermocouple within Test Specimen


Thermocouple x (cm) y (cm) z (cm) Identification
Concrete layer 1 81 66 0.0 T05
Concrete layer 2 66 54 0.0 T20
Concrete layer 3 52 38 0.0 T08
Concrete layer 4 35 25 0.0 T03
Concrete layer 5 17 13 0.0 T07
Reinforcement bar 1 18 55 3.1 T13
Reinforcement bar 2 29 44 3.1 T12
Reinforcement bar 3 57 34 3.1 T11
Reinforcement bar 4 68 24 3.1 T14
Timber beam 1 10 54 8.0 T09
Timber beam 2 35 54 5.0 T01
Timber beam 3 80 54 5.0 T02
Timber beam 4 35 16 5.0 T16
Timber beam 5 55 16 5.0 T19
Timber beam 6 90 16 8.0 T24
Timber groove 1 47 57 7.0 T04
Timber groove 2 73 50 7.0 T10
Timber groove 3 50 19 7.0 T17
Timber groove 4 50 13 7.0 T18
Open pipe 1 18 33 4.9 T25 Fig. 4. Test observations: (a) untested specimen; (b) cracks occurred
Open pipe 2 15 4 3.3 T23
after 15 min of fire exposure (picture taken at the end of test); (c) water
Open pipe 3 63 33 4.0 T30
Open pipe 4 76 4 5.4 T22 from the cracks in concrete slab after 15 min; (d) steam evaporating
Closed pipe 1 22 69 5.0 T27 though openings; (e) ignition of timber beams after 49 min; (f) no spal-
Closed pipe 2 24 43 3.2 T29 ling at bottom surface of the concrete slab at the end of the test;
Closed pipe 3 72 69 3.5 T21 (g) melted pipe (thermocouple highlighted); and (h) char layer at
Closed pipe 4 70 43 5.0 T28 the bottom surface (thermocouple highlighted)

© ASCE A4014009-4 J. Perform. Constr. Facil.

J. Perform. Constr. Facil., 2014, 28(6): A4014009


device. The novel TCC system was therefore tested with the con- beams [Fig. 4(h)] became visible. A char layer of nearly 6 mm
crete slab exposed to the fire, neglecting the possibility of fire ex- was measured. As the ignition occurred during minute 49; it was
posure from other sides. To prevent such a fire exposure in practice, determined that the rate of burning of wood was approximately
constructive measures have to be used. The practical application of 0.5 mm=min. The nominal burning rate for the Glulam timber is
the system was investigated in a separate work package of the 0.7 mm=min (SIA 2004) when one of the sides is directly exposed
Swiss wood-concrete deck project. to fire. In this particular case the timber beam was not exposed to
fire directly, therefore the burning rate was less than the standard
Test Results nominal burning rate. The thickness of the char layer can be calcu-
The subsequent observations were made during the experiment. lated by dchar ¼ β 0 t (mm=min), where t is the time after ignition
After 15 min, thermal cracks appeared on the upper side of the until the end of the process; and β 0 is the rate of burning of
concrete slab [Fig. 4(b)] and water from the pipes leaked trough the wood.
these cracks [Fig. 4(c)]. This implies that the multilayer pipes Fig. 5(a) shows the temperature profiles obtained from the oven,
melted before minute 15 and suggests that the pipes cannot resist which verifies that the fire conditions within the oven is in accor-
Downloaded from ascelibrary.org by University of Waterloo on 09/03/21. Copyright ASCE. For personal use only; all rights reserved.

a temperature above 95°C (as specified in the manufacturer’s spec- dance with the STtC pattern. Fig. 5 also shows the temperature pro-
ifications). After 16 min, the water in the open pipe was boiling files obtained from the reinforcement bars, the open and closed
and steam was evaporating though the openings [Fig. 4(d)]. After pipes, at the contact surface of the timber beam, and at the middle
35 min, the leaked water which collected on the upper side of the of the penetration depth. Fig. 5(b), upper curves, shows that the
slab started boiling and evaporating. After 49 min, smoke was temperature variation at the concrete surface exposed to the fire
observed, which implies that the timber beams started charring conditions was not in accordance with the exact STtC profile be-
[Fig. 4(e)]. After 60 min, the exposed concrete surface showed cause there is continuous heat transfer between the surface exposed
no damage. There were neither cracks nor was there any visible to fire and the inner part of the concrete slab. Fig. 5(b), lower
spalling [Fig. 4(f)]. curves, shows the temperature variation in the reinforcement bars
After separating the concrete from the timber beams, the melted during the experiment. The profiles show a linear variation of
pipes [Fig. 4(g)] and the charred bottom surface of the timber temperature in reinforcement bars and the maximum temperature

(a) (b)

(c) (d)

Fig. 5. Rate of change of temperature: (a) within the oven; (b) at concrete (upper curves) and reinforcement bars (lower curves); (c) timber bottom
surface (upper curves) and at grooves (lower curves); and (d) pipes (curves for open and closed pipes shown together)

© ASCE A4014009-5 J. Perform. Constr. Facil.

J. Perform. Constr. Facil., 2014, 28(6): A4014009


Table 3. Properties of Different Materials for Thermal Analysis
Material ρ (kg=m3 ) λ (W=m-K) c (J=kg-K)
Concrete 2,300 — —
Wood 450 — —
Steel 7,850 — —
Water 1,000 0.60 4,186
Multilayer pipe 2,700 0.40 2,300

was nearly 450°C after 60 min. Fig. 5(c), upper curves, shows the
temperature profiles for the timber beam at contact surface. The
maximum temperature does not exceed a value of 400°C that
did not show any sever effects on the TCC system during the
Downloaded from ascelibrary.org by University of Waterloo on 09/03/21. Copyright ASCE. For personal use only; all rights reserved.

observation. Temperature profiles for both pipes are fairly similar


on both the top and bottom sides of the pipes [Fig. 5(d)]. Conse-
quently, in the subsequent numerical model no distinction was
made between open and closed pipe. The pipes started melting after
12 min at a temperature of nearly 100°C. Right after this occurrence
the water started boiling. Fig. 6. Isometric view of numerical model

Numerical Simulation Temperature Profiles


The secondary objective of the research reported in this paper was The modeled rates of change of temperature for the different
to generate and validate a numerical model to provide a basis for materials (Fig. 7) illustrate that the temperature increases nearly
parameter optimization and predicting fire resistance ratings of linearly for all the materials with respect to time. Initially, the
optimized systems. temperature of the reinforcement bars and the pipes increased at
a higher rate initially, while the temperature in the grooves in-
creased at a slightly slower rate. After approximately 20 min, all
Thermal Material Properties
further increases in temperature were almost perfectly linear with
To model the behavior of the novel TCC system under fire condi- time. The maximum temperature in reinforcing steel bars is nearly
tion, the thermal conductivity and specific heat are two core proper- 480°C after 60 min for 25 mm concrete cover to the reinforcement.
ties, which have to be precisely taken into consideration for thermal This finding implies that the steel retains approximately 80% of its
analysis of the structural element. Thermal conductivity λ is an strength (Lie 1978). The maximum temperature at the bottom sur-
intrinsic material property (independent of material size, shape or face of the pipe is nearly 100°C after 10 min. The maximum tem-
orientation), which describes the material’s ability to conduct heat. perature in the pipe is more than 400°C after 60 min, which implies
The specific heat capacity c of a solid or liquid is defined as the heat that there will be no water at the end of the process and the pipe will
required to change the temperature of a unit mass of substance by have melted. The temperature in the timber beam on the bottom
1°C. Thermal properties of these materials are also temperature surface is approximately 360°C after 60 min. Ignition may occur
dependent but due to unavailability of the required data, data at after 48 min when the temperature on the bottom of the timber
ambient temperature were used. Table 3 shows the properties used beam reaches about 300°C. Therefore the thickness of the char
in the transient thermal analysis to determine the temperature at layer can be calculated as dchar ∼ β 0 × t ¼ 0.5 × 12 ¼ 6 mm, with
different times and locations using the ISO 834-1 (ISO 1999) STtC; t being the time from ignition until the required (R60) fire resistant
thermal conductivity and specific heat of different materials for the rating; and β 0 being the rate of burning of wood (0.5 mm=min in
purpose of modeling were taken from the literature (Bailey 2005). this paper, from the experimental investigations). Roughly 6 mm of
the beam will be converted into char layer. The remaining section of
Transient Thermal Analysis the beam is able to sustain the weight of the slab and the timber
beam will not lose its connection with the concrete slab.
The commercial software package Workbench (Ansys 2011) was
used; Fig. 6 shows an isometric view of the model, exploiting
symmetry. Only a portion of the test specimen was modeled as the Discussion
temperature variation along the bottom surface of the system was
nearly identical during the fire tests (Fig. 5). This modeled section Fig. 7 compares the temperature profiles obtained from the exper-
of the system included all components and materials of the TCC imental and numerical investigations for the different materials.
system. SOLID87 elements (three-dimensional, tetrahedral, 10 All the observations and results are used to define the effective fire
node) were deemed most appropriate for all the materials which resistance rating of the TCC system. Temperature profiles from
for a thermal analysis with free mesh option. Small elements thermocouples are compared with the temperature profile from the
(10 mm in size) were used to refine the mesh at the contact surface thermal analysis for different materials and locations. The results
of the timber beam and the concrete slab. Otherwise, the model obtained from the experimental procedure and the transient thermal
relied on the automatic modeling procedures provided by Work- analysis show adequate resemblance.
bench. Transient thermal analyses determined thermal quantities Thermal cracks occurred due to thermal expansion of the slab
that varied over time. A transient thermal analysis can be either because of excessive temperature difference within the slab and its
linear or nonlinear. Since thermal properties of most materials such surroundings. The temperature of the slab at the surface that was
as thermal conductivity, specific heat, or density vary with temper- exposed to fire was more than 700°C after 15 min. No spalling in-
ature, the analysis had to be nonlinear. stabilities occurred during the entire process, which is one of the

© ASCE A4014009-6 J. Perform. Constr. Facil.

J. Perform. Constr. Facil., 2014, 28(6): A4014009


Downloaded from ascelibrary.org by University of Waterloo on 09/03/21. Copyright ASCE. For personal use only; all rights reserved.

Fig. 7. Comparison of the experimental and numerical results for temperatures of the different materials when the bottom surface of the concrete slab
exposed to the ISO 834-1 (ISO 1999) STtC

most critical observations in concrete materials when subjected to efforts, ideally validated by more tests, on the fire performance of a
high temperatures. complete slab that also consider live loads are recommended.
The temperature in the reinforcement bars during the experi-
ments was nearly 450°C after 60 min, which is relatively compa-
rable to the temperature obtained from the thermal simulation. At Conclusions
this temperature, the reinforcement bars retain 90% of their initial
strength (Lie 1978), which is the most important characteristic re- The research reported in this paper allow the following main con-
quired to keep the concrete slab cohesive even after the cracks. clusions to be drawn: (1) the effective fire resistance rating of the
In the fire test, water from the pipes appeared on the top of the TCC system can be assumed to be 60 min (R60), (2) the minimum
concrete slab during minute 15 through the cracks in the slab. concrete cover of 25 mm for the reinforcement is sufficient for R60
fire resistance rating of TCC system, (3) as the results for open and
Analysis showed that the boiling process of water and the melting
closed pipe did not show any difference it can be concluded that
process of pipes occurred between minutes 12 and 15 of the pro-
there is no risk of pipes bursting if the water is trapped within the
cess. These results were verified from the numerical transient ther-
pipes (after the pipe melts, the water will come out of the thermal
mal analysis, which showed that the water within the pipes reached
cracks in the concrete slab), (4) concrete cover of 50 mm is suffi-
an average temperature of 100°C after 12 min. In addition, this pro-
cient for the timber beam grooves to sustain weight of the concrete
cess confirmed the maximum operating temperature for the multi-
slab for one hour of fire exposure without losing the beam-slab con-
layer pipes (95°C) given by the company as the water must have
nection, and (5) the validated numerical model can be used to ana-
leaked from the melted pipes.
lyze the temperature profiles for different setups to allow for system
Furthermore, ignition in the timber beams was observed by
optimization.
smoke, which occurred after 49 min at a temperature of close to
300°C. In the numerical model, this temperature at the bottom
of the beam was observed after 48 min, which verifies the obser- Acknowledgments
vations from experimental investigation. From the 30-mm beam
inside the concrete slab, almost 6 mm were converted into char The research reported in this paper was funded by the Swiss
at the end of the process, yet the residual height of the beams sus- Commission for Technology and Innovation (CTI) and carried
tained the weight of the concrete slab. The experimentally observed out in collaboration with Renggli AG, Pirmin Jung Ingenieure
thickness of the char layer and the thickness as calculated based on für Holzbau AG, Makiol+Wiederkehr Holzbau-Ingenieure, Jomos
the numerical results are in excellent agreement. Further modeling Gruppe, and Rekag AG. The help of the technicians at the Tech

© ASCE A4014009-7 J. Perform. Constr. Facil.

J. Perform. Constr. Facil., 2014, 28(6): A4014009


Park of the Bern University of Applied Sciences in Biel and of Gosselin, G. C., and Lie, T. T. (1987). “Provision of fire resistance:
the fire lab at ETH Zurich is kindly acknowledged. The support Evolution of design approaches.” Proc., Canadian Society for Civil
by Meng students Robbe Drugmand and Ilana Danzig at the Engineering (CSCE) Centennial Conf., National Research Council,
University of British Columbia is also appreciated. Ottawa, 3–19.
HakaGerodur. (2011). “Product specifications.” 〈http://www.hakagerodur
.ch〉 (May 22, 2014).
Harmathy, T. Z., Sultan, M. A., and MacLaurin, J. W. (1987). “Comparison
References of severity of exposure in ASTM E119 and ISO 834 fire resistance
tests.” J. Test. Eval., 1, 371–375.
Ansys. (2011). “Documentation for ANSYS. Release 11.” ANSYS, Inc., Hehl, S., Meena, R., Tannert, T., and Vallee, T. (2013). “Experimental
Canonsburg, PA, 〈http://www.ansys.com〉 (May 22, 2014). and numerical investigations on groove connections for novel timber-
ASTM. (2003). “Standard test methods for fire tests of building construc- concrete-composite system.” J. Perform. Constr. Facil., 10.1061/
tion and materials.” ASTM E119, West Conshohocken, PA. (ASCE)CF.1943-5509.0000549, A4014010.
Babrauskas, V. (2002). “Ignition of wood–A review of the state of the art.” Negrão, J., Leitão de Oliveira, C., Maia de Oliveira, F., and Cachim, P.
J. Fire Prot. Eng., 12(3), 163–189. (2010). “Glued composite timber-concrete beams. I: Interlayer connec-
Downloaded from ascelibrary.org by University of Waterloo on 09/03/21. Copyright ASCE. For personal use only; all rights reserved.

Bailey, C. (2005). “One stop shop in structural fire engineering.” 〈http:// tion specimen tests.” J. Struct. Eng., 10.1061/(ASCE)ST.1943-541X
www.structuralfiresafety.org〉 (May 22, 2014). .0000228, 1236–1245.
Branthwaite, K. (2010). “A finite element analysis of the fire resistance of ISO. (1999). “Fire-resistance tests–Elements of building construction–Part
FRP reinforced member.” Research Rep. Prepared for the School of 1: General requirements.” ISO 834-1, Geneva.
Civil Engineering, Univ. of Leeds, Leeds, U.K. Kodur, V. K. R., Wang, T. C., and Cheng, F. P. (2004). “Predicting the fire
Brunner, M., Romer, M., and Schnüriger, M. (2006). “Timber-concrete- resistance behaviour of high strength concrete columns.” Cement
composite with an adhesive connector (wet on wet process).” Mater. Concrete Compos., 26(2), 141–153.
Struct., 40(1), 119–126. Lane, W. (2005). “Ignition, charring and structural performance of
Clouston, P., Bathon, L., and Schreyer, A. (2005). “Shear and bending per- laminated veneer.” Fire Engineering Research Rep. 05/03, Univ. of
formance of a novel wood-concrete composite system.” J. Struct. Eng., Canterbury, Christchurch, New Zealand.
10.1061/(ASCE)0733-9445(2005)131:9(1404), 1404–1412. Lie, T. T. (1978). “Calculation of the fire resistance of composite concrete
Clouston, P., and Schreyer, A. (2008). “Design and use of wood-concrete floors and roof slabs.” Fire Technol., 14(1), 28–45.
composites.” Pract. Period. Struct. Des. Constr., 10.1061/(ASCE) National Research Council (NRC). (2010). National building code of
1084-0680(2008)13:4(167), 167–174. Canada. Fire and structural protection of buildings, Ottawa.
European Committee for Standardization (CEN). (2004). “Eurocode 5: Natterer, J. K. (2002). “New technologies for engineered timber structures.”
Design of timber structures–Part 1-1: General–Common rules and rules Prog. Struct. Eng. Mater., 4(3), 245–263.
for buildings.” EN 1995-1-1:2004, Brussels, Belgium. O’Neill, J. W. (2009). “The fire performance of timber-concrete composite
Fletcher, I. A., Welch, S., Torero, J. L., Carvel, R. O., and Usmani, A. floors.” MASc. thesis, Univ. of Canterbury, Christchurch, New Zealand.
(2007). The behaviour of concrete structures in fire, Building Research Peng, L. (2010). “Performance of heavy timber connections in fire.” Ph.D.
Establishment (BRE) Center for Fire Safety Engineering, Edinburgh, thesis, Carleton Univ., Ottawa.
U.K. Swiss Society of Engineers and Architects (SIA). (2004). “Timber struc-
Fragiacomo, M. (2005). “A finite element model for long-term analysis of tures.” SIA 265, Zurich, Switzerland.
timber-concrete composite beams.” Struct. Eng. Mech., 20(2), 173–189. Tsai, K., Carradine, D., Moss, P., and Buchanan, A. (2013). “Charring rates
Fragiacomo, M. (2006). “Long-term behaviour of timber-concrete for double beams made from laminated veneer lumber (LVL).” Fire
composite beams. II: Numerical analysis and simplified evaluation.” Mater., 37(1), 58–74.
J. Struct. Eng., 10.1061/(ASCE)0733-9445(2006)132:1(23), 23–33. White, R. H. (2000). “Charring rate of composite timber products.”
Frangi, A., Erchinger, C., and Fontana, M. (2008). “Charring model for Proc., Int. Conf. Wood and Fire Safety, Tech. Univ. of Zvolen, Slovak
timber frame floor assemblies with void cavities.” Fire Saf. J., 43(8), Republic, 353–363.
551–564. Yeoh, D., Fragiacomo, M., De Franceschi, M., and Heng Boon, K. (2011).
Frangi, A., Fontana, M., Hugi, E., and Jöbst, R. (2009). “Experimental “State of the art on timber- concrete composite structures: Literature
analysis of cross-laminated timber panels in fire.” Fire Saf. J., 44(8), review.” J. Struct. Eng., 10.1061/(ASCE)ST.1943-541X.0000353,
1078–1087. 1085–1095.
Frangi, A., Knobloch, M., and Fontana, M. (2010). “Fire design of timber- Zadrazil, T., Vodak, F., and Kapickova, O. (2004). “Effect of temperature
concrete composite slabs with screwed connections.” J. Struct. Eng., and age of concrete on strength- porosity relation.” Acta Polytech.,
10.1061/(ASCE)ST.1943-541X.0000101, 219–228. 44(1), 53–56.

© ASCE A4014009-8 J. Perform. Constr. Facil.

J. Perform. Constr. Facil., 2014, 28(6): A4014009

You might also like