Waghner C. Rocha, José J.R. Silva, Tiago A. Pires and Leonardo M. Costa
An increase in ^ the axial restraining levei or in load levei reduced the criticai time and
temperature of the^columns. However the influence of these parameters was not significant-fo~r
the fire behavior of the columns. Ali tests presented less than 10 min for the criticai "time "and
criticai temperatures between 500°C and 700°C. This criticai time is lower than th^minimal"flre
resistance required by the NBR 14432:2001 that is 30min.
IFireSS - International Fire Safety Symposium
Coimbra, Portugal, 20th-22"d Ãpril 2015
FIRÍJÍE^TION?. F CONCRETEWITH AND WITHOUT PP FIBRES:
EXPERIMENTAL
ANALYSIS AND
NUMERÍCÀLÍrM
ULATION"""
4. CONCLUSIONS
TNS.Paper_presenteda series of 8 fire tests on cold formed columns developed in laboratory of
structuresand materials of Federal universlty °f Pernambuco. The main"conclusÍons"of'íhïs
research were:
' The ^iÍica'time. oMhe tested_columns were less than 10min and criticai temperature
was between 500°C and 700°C;
. The higher load levei or axial restraining to thermal elongation, the lower is the fire
resjstance ofthe columns;
' ^B^So'Od(t3hO'nÏm"' are lowe> 'han 'he minim" r"° re"SBnce ^re'ented
5. AKNOWLEDGMENTS
LO
F.NWMCIdueJhe.
p;qlect
"planeiamento de Emergência de Complexos Industriais e noseu Entorno" n° 01. 08. 0618. 03 that
support7he"devdopmaento'f°th^
re^a^hA
US '"UUSInals e no6. MAIN REFERENCES
[1]
[2]
[3]
pire.sLT.A'c" Rodrigues-.. J-p'c' & Rê9° s"va' JJ- - Hre Resistance of concrete filled
CJK
U!^ho"ow.
col.
umns with ^trainedthermal, Journal ofConstructionalsteemesearch.
., p. 82-94. ----. -.,,
(:orre'a;A:J-M'c-; &,Rodri9ues, J.P.C. - Fire Resistance of steel columns with restrained
tion, Fire Safety Journal, 50, 2012, p. 1-11.
Almejda'-s, j'c' ^Análise do comportamento a temperaturas elevadas de elementos de
arço. formados a. friocomPrimidos considerando restrição ao alongamento"térmi'co^
Jhesis, University of Sao_Paulo - USP - Brazil, 2012, 292 p. (in portuguease)'"'""""'
[4]Íam;.. L;-Rodri9ues'J'. p'c;;. & &lva' Ls- - Experimenta^nalysisTn^-formed steel
beams subjected to fire, Thin-Walled Structures, 74, 2014, p. 104--1l7
Paulo Piloto6
Professor
Poly. Inst. Bragança
Portugal
Luís M. R. Mesquita
Professor
Poly. Inst. Bragança
Portugal
Carlos Balsa
Professor
Poly. Inst. Bragança
Portugal
Keywords: Fire reaction; Concrete; PP fibres; Thermal performance; heat release rate.
1. INTRODUCTION
Sï, ïal dements-ofre'nforced concrete-in generai- Present 9°°d performance in case of fíre.
Hnole
;eLmorlrecenLstructures.
have adapted newt^es-
^^^S
s;reang
ü's^-^SÏg. Jte:Lpresenting Ïfferent thermo me^hanical behaviour:"acquiringa"sp^al
;^<^^^'^ï:9^:e^e^°^:p^e^:^:ni^^
Ï. add!t'^tí^ypropylen<flbres. (pp) to n^l"cx>mpone^7^du^g"t^ ^^n^>S
pressure of the material through the channels created by the fusion of thefibres.'
^edeTspeh^ean;lÏd^ï:lT^de;fírc^nd'tion.
san^thedevelopment
of new numen^'^de's^has_a"owed.
theassessment of more or less^Plex'phenomena"tol
ïdeute'rcm'^
temperature evolution and other state variables, enabling
differentle^s'of"app°rol
achueT'u's;^
^pmd"ls wt)lw ~Departmel" °fAPP"ed Mecl'amcs-POMCChBic Inslïule ofBraEM'a- campus s^ Apolonia, ap. 1134, 5301-857 B.a..nca.
Paulo Piloto, Luís M. R. Mesquita and Carlos Bolsa
coupled or uncoupled field interaction (thermal, mechanical, hydrodynamic, chemical). This
investigation studies the thermal performance of a two dimensional model, using nonlinear and
transient finite element analysis.
2. MATERIAL AND METHODS
The experimental analysis is based on calorimetric test using small-scale samples of concrete, with dimensions 100x100x40 mm using the test method EN ISO 13927 [1]. In addition to the standard test, normally used to determine the mass loss rate and the heat relesse rate, four thermocouples type k were positioned in different coordinates of the samples ana the infrared thermography camera (FLIR 365) was used to evaluate temperature field in one lateral surface of each sample. A numerical simulation model was defined to evaluate the thermal performance
of samples when submitted to different heat fluxes 35 and 75 [kW/m2] and different levei of
fibres contents, see table 1. In arder to use the cone calorimeter experiments to validate the numerical model, small changes to standard test procedure were made.
Table 1: Tested samples.
Samples Materiais
Heat flux[kW/m^]
01 02 03 BF1 600 BF2 600 BF1 1200 BF2 1200
AN AN AN
AN + PP [600 g/m3]
AN + PP [600 g/m3]
AN+PP[1200g/m3]
AN+PP[1200g/m3]
35 35 75 35 75 35
75
The thermal performance of the samples depends on the thermal balance in the boundaries.
The Eq. 1 should be solved in the two dimensional domain of the sample (fí), taking into
account the exchange of heat with the surroundings (ô0;). On the top surface of the sample
(3Qi), the heat flow balance (input and output) should verify Eq. 2. The net incident heat flux at
the top surface of the sample is composed by the heat flux coming from the cone heater,
radiation reflected from the surface, convective and radiative heat lesses. On the lateral
surfaces ofthe sample (002) the convective and radiative heat losses should be considered by
Eq. 3 and at the bottom surface of the domain (õQs) the adiabatic condition may be assumed,
Eq. 4. In these equations T represents the main state variable (temperature), the thermal
properties of concrete are represented by the specific mass P(J), specific heat CP(T-) and
conductivity /I(T-). The emissivity of concrete s was considered equal to 0.7. The
Stefan-Boltzmann coefficient is denoted by o- and h represents the convection coefficient that was
approximated by an experimental correlation for a hot horizontal plate in air, with the hot surface
Paulo Piloto, Luís M. R. Mesquita and Carlos Balsa
uppermost [2]. Tg represents the ambient temperature. These equations should consider the
nonlinear behaviour of material properties [3]. The balance model is represented in figure 1, as
well as the setup used in experiments.
V. (^T). VT) p(T)-Cp(T)-ST/Bt
(Q)
(^). VT). n=qo-/7(7--Ta)-aT(T4-Ta4)-(l-f)90
(3Ql)
(Â(T). Vr). n=-/7(T-7a)-aT(7-4-Ta4) (5^2)
(A(D. VT). n=0
(003)
(1)
(2)
(3)
(4)
^?Rn
»fl (tf (t( ( n..tó
|W-7;y ^U" - :c«(T'-ffl
lr^
a) Thermal model. b) Cone heater and thermocouples. Figure : Test model and instrumentation.
3. RESULTS
Figure 2 presents the thermocouple measurements during tests. These values are useful to
validate the numerical thermal model. The heat release rate was also measured and there was
no significant difference between samples with and without PP fibres for the heat flux levei of 35
Paulo Piloto, Luís M. R. Mesquita and Carlos Bolsa
coupled or uncoupled field interaction (thermal, mechanical, hydrodynamic, chemical). This
investigation studies the thermal performance of a two dimensional model, using nonlinear and
transient finite element analysis.
2. MATERIAL AND METHODS
The experimental analysis is based on calorimetric test using small-scale samples of concrete, with dimensions 100x100x40 mm using the test method EN ISO 13927 [1]. In addition to the standard test, normally used to determine the mass loss rate and the heat relesse rate, four thermocouples type k were positioned in different coordinates of the samples ana the infrared thermography camera (FLIR 365) was used to evaluate temperature field in one lateral surface of each sample. A numerical simulation model was defined to evaluate the thermal performance
of samples when submitted to different heat fluxes 35 and 75 [kW/m2] and different levei of
fibres contents, see table 1. In arder to use the cone calorimeter experiments to validate the numerical model, small changes to standard test procedure were made.
Table 1: Tested samples.
Samples Materiais
Heat flux[kW/m^]
01 02 03 BF1 600 BF2 600 BF1 1200 BF2 1200
AN AN AN
AN + PP [600 g/m3]
AN + PP [600 g/m3]
AN+PP[1200g/m3]
AN+PP[1200g/m3]
35 35 75 35 75 35
75
The thermal performance of the samples depends on the thermal balance in the boundaries.
The Eq. 1 should be solved in the two dimensional domain of the sample (fí), taking into
account the exchange of heat with the surroundings (ô0;). On the top surface of the sample
(3Qi), the heat flow balance (input and output) should verify Eq. 2. The net incident heat flux at
the top surface of the sample is composed by the heat flux coming from the cone heater,
radiation reflected from the surface, convective and radiative heat lesses. On the lateral
surfaces ofthe sample (002) the convective and radiative heat losses should be considered by
Eq. 3 and at the bottom surface of the domain (õQs) the adiabatic condition may be assumed,
Eq. 4. In these equations T represents the main state variable (temperature), the thermal
properties of concrete are represented by the specific mass P(J), specific heat CP(T-) and
conductivity /I(T-). The emissivity of concrete s was considered equal to 0.7. The
Stefan-Boltzmann coefficient is denoted by o- and h represents the convection coefficient that was
approximated by an experimental correlation for a hot horizontal plate in air, with the hot surface
Paulo Piloto, Luís M. R. Mesquita and Carlos Balsa
uppermost [2]. Tg represents the ambient temperature. These equations should consider the
nonlinear behaviour of material properties [3]. The balance model is represented in figure 1, as
well as the setup used in experiments.
V. (^T). VT) p(T)-Cp(T)-ST/Bt
(Q)
(^). VT). n=qo-/7(7--Ta)-aT(T4-Ta4)-(l-f)90
(3Ql)
(Â(T). Vr). n=-/7(T-7a)-aT(7-4-Ta4) (5^2)
(A(D. VT). n=0
(003)
(1)
(2)
(3)
(4)
^?Rn
»fl (tf (t( ( n..tó
|W-7;y ^U" - :c«(T'-ffl
lr^
a) Thermal model. b) Cone heater and thermocouples. Figure : Test model and instrumentation.
3. RESULTS
Figure 2 presents the thermocouple measurements during tests. These values are useful to
validate the numerical thermal model. The heat release rate was also measured and there was
no significant difference between samples with and without PP fibres for the heat flux levei of 35
Paulo Piloto, Lias M. R. Mesquita and Carlos Balsa
.'[. 11 -l
- -r n l: ni l: n.' -r. n --r', li: -;-i.ni--1^'.; tí( ! < l, » l' Hl i (.'il- .."' I \ M } M'-! "". Í . W i^---'
RFt i'nc I;HIt l^y< -14hit t'**'1 > 11 \\^ ^ \\s-<i>. !. ^^<. ^^ O i- \\^s
flRR(-kWm'i -«í -D: -BrÉ f')0 -tíFJ !2<H'- -03 ~ "tít'2 L>UÜ - -BI 2 12t)n
,
" u
'l
^
"" ; - ^'7':.
;lsÍ's
:,^s;sSS sw^,
;-y^í^^'^^-1"-''-\ . -'...'~|-.ï.
'.^f-SSS
^^A. ^^ï'-';. -^'
1:0-) ) siri 2-. W ^0(ï')
11-1
Figure 2: Thermal performance with simulation results and hleat Release Rate ofsame
samples.
IFireSS International Fire Safety Symposium
Coimbra, Portugal, 20th-22"d April 2015
EXPERIMENTAL INVESTIGATION ON THE BEHAVIOUR OF
COLD-FORMED STEEL COLUMNS SUBJECTED TO FIRE
4. CONCLUSIONS
The thermal behaviour of concrete samples, with and without polypropylene fibres was
presented, when submitted to a uniform heat fíux of 35 and 75 [kW/m ]. This study was
performed experimentally with two methods and numerically, with the validation ofthe numerical model. The numerical model hás the limitation of predicting the motion of the moisture in the samples.
It can also be concluded that addition of different amounts of polypropylene fibres hás no significant effect on the heat release rate and on the thermal performance of the samples.
5. REFERENCES
[1] CEN, EN ISO 13927. Plastics - Simple heat release test using a conical radiant heater and a thermopile detector. Brussels: CEN - Comité Européen de Normalisation." 2003. [2] Staggs, J. E. J. ; Whiteley, R. H. Modelling the combustion of solid-phase fuels in cone
calohmeter experíments. Fire and Materiais, vol. 23, issue 2, 1999, p. 63-69.
[3] CEN EN 199212. Eurocode 2: Design of concrete structures Part 12: General rules -Structural fire design. Brussels, December 2004.
Hélder D. Craveiro PhD student
University of
Coimbra
João P. Rodrigues
Professor
University of
Coimbra
Luís Laím Researcher
University of
Coimbra
Keywords: cold-formed steel, column, restraining, buckling
1. INTRODUCTION
The market share of cold-formed structural steelwork keeps growing over the past decade, especially in low rise residential, industrial and commercial buildings. This led to an understanding on how the fire safety design for cold-formed steel (CFS) structures is important. However, só far, most research hás been focusing on hot-rolled steel members which present significant differences when compared with cold-formed steel members, due to their manufacturing processes and geometry. It is clear that accurate design rules under fire conditions for cold-formed steel members are needed. The EN 1993-1-2:2005 [1] predicts that the design methodology for hot-rolled steel members is also applicable to cold-formed steel members with class 4 cross-sections, establishing the same reduction factors for the yield strength ofthe steel and limiting the criticai temperature to 350°C. The investigation carried out in the field hás shown that these design guidelines are not accurate [2-5]. Só far the research conducted hás been focusing on the individual buckling phenomena such as local, distortional ana global [5-7]. Moreover the research carried out does not consider the influence of restraint
to thermal elongation on the structural behavior of CFS columns. However this type of
Correspndent Author - Departamento de Engenliaria Civil da Faculdade de Ciências e Tecnologia da Universidade de Coimbra. Rua Luís Reis Santos. Polo
