-
ARTICLE IN PRESS
Fire Safety Journal 44 (2009) 376386Contents lists available at
ScienceDirectFire Safety Journal0379-71
doi:10.1
CorrE-m
yong.wajournal homepage: www.elsevier.com/locate/firesafEffects
of partial fire protection on temperature developments in steel
jointsprotected by intumescent coatingX.H. Dai, Y.C. Wang , C.G.
Bailey
School of Mechanical, Aerospace and Civil Engineering,
University of Manchester, PO Box 88, Manchester M60 1QD, UKa r t i
c l e i n f o
Article history:
Received 21 May 2008
Received in revised form
19 August 2008
Accepted 25 August 2008Available online 2 October 2008
Keywords:
Steelconcrete composite joints
Partial fire protection
Intumescent coating
Temperature
Fire tests
Fire resistance
Unprotected bolts12/$ - see front matter & 2008 Elsevier
Ltd. A
016/j.firesaf.2008.08.005
esponding author. Tel.: +441613068968.
ail addresses: [email protected],
[email protected] (Y.C. Wang).a b s t r a c t
This paper presents experimental results of temperature
distribution in fire in four typical types of
steelconcrete composite joint (web cleat, fin plate, flush
endplate and flexible endplate) with different
fire-protection schemes. The test specimens were unloaded and
the steelwork of each joint assembly
was exposed to a standard fire condition [ISO 834, 1975: Fire
Resistance Tests, Elements of Building
Construction, International Organization for Standardization,
Geneva] in a furnace. In total, 14 tests
were conducted, including 4 tests without any fire protection
and 10 tests with different schemes of fire
protection. The main objective of these tests was to investigate
the effects of three practical fire-
protection schemes as alternatives to full fire protection of
the entire joint assembly. The three
alternative methods of fire protection were: (1) protecting a
segment, instead of the entire length, of the
beams; (2) unprotected bolts and (3) protecting the columns
only. The main results of these tests are:
(1) if all the steel work (excluding the bolts) in the joint
assembly was protected, whether or not
protecting the bolts had very little effect on temperatures in
the protected steelwork other than the
bolts. The bolt temperatures were higher if they were not
protected than if they were protected, but the
unprotected bolt temperatures in a joint with fire protection to
other steelwork were much lower than
bolt temperatures in a totally unprotected joint; (2) as far as
joint temperatures are concerned,
protecting a segment of 400mm of the beam was sufficient to
achieve full protection and (3) if only the
column was protected, only the joint components that were in the
immediate vicinity of the column
(such as welds) developed noticeably lower temperatures than if
the joint assembly was unprotected,
but due to heat conduction from the unprotected steel beams,
these temperature values were much
higher than if the joint assembly was protected. Furthermore,
the column temperatures in the joint
region were much higher than the protected column
temperatures.
& 2008 Elsevier Ltd. All rights reserved.1. Introduction
Joints are critical members in steel-framed structures.
Inparticular, how joints behave in steel-framed structure has
acritical influence in controlling progressive collapse of
thestructure under accidental fire attack. Despite extensive
previousresearch on steel-framed structures in fire, which has
resulted inthe development of fire engineering design methods that
are nowbeing routinely adopted in steel structural design, large
gaps stillexist in the understanding of joint behaviour in fire.
Following theWorld Trade Center disaster, a number of authoritative
organisa-tions [1,2] have identified joint integrity as key to
maintainingstructural integrity in a fire and have called for
extensive researchon joints under fire conditions.ll rights
reserved.The current practice [3] to ensure that joints have
sufficient fireresistance is simple: to protect joints to the
highest level of fireprotection based on the connected members.
This is based on theassumption that because joint components have
lower sectionfactors compared to the connecting members (defined as
the ratioof the fire-exposed surface area to the volume of steel
beingheated) the temperature rise in the joint components is
lowerthan in the connected members. One immediate shortcoming
ofthis approach is that some joint components may be subject
tohigher levels of loading than the connected members.
Moreimportantly, under fire conditions, the behaviour of a
steelstructure is complex with forces in different members
changingduring the entire course of fire exposure. These forces
aretransmitted from one connected member to another,
mainlydependent on the behaviour and performance of the
joints,making understanding joint behaviours in fire a key factor
instructural fire design.
Understanding structural behaviour in fire involves threegeneral
steps: quantifying the fire behaviour, assessment of
www.sciencedirect.com/science/journal/fisjwww.elsevier.com/locate/firesafdx.doi.org/10.1016/j.firesaf.2008.08.005mailto:[email protected],mailto:[email protected]
FaisalHighlight
Arbab FaisalHighlight
Arbab FaisalSticky Notewhy study fire behavior of joints
-
ARTICLE IN PRESS
X.H. Dai et al. / Fire Safety Journal 44 (2009) 376386
377temperature development in the structure and understanding
thestructural behaviour at elevated temperatures. For joints,
thefocus has mainly been on joint structural behaviour at
elevatedtemperatures [412]. This paper deals with quantification of
thetemperature development in different joint components.Although
there have been studies of temperature developmentin steel joints
in fire [1316], these studies were concerned withthe temperatures
in unprotected steel joints. The focus of thispaper is on protected
steel joints. In particular, this paper dealswith temperatures in
joints protected by intumescent coating.Steel being a thermally
high conductive material, the temperaturerise in unprotected steel
exposed to fire attack is quick, resultingin rapid loss in strength
and stiffness of steel in fire. To ensuresufficient fire resistance
of steel-framed structures, fire protectionis often required to
limit temperature rises in steel. Currently,intumescent coating is
the most popular type of fire protection,representing about 50% of
the passive fire-protection market inthe UK. Application of
intumescent coating fire protection tojoints can be a
time-consuming process. Therefore, an importantassociated objective
of this study is to develop rules for effectiveapplication of
intumescent coating in practice. Intumescentcoating may be applied
either on- or off-site. In on-site applicationof intumescent
coating, unprotected steel members with suitableconnection
components are assembled in the fabrication shop andapplication of
the fire protection commences after site erection ofthe steel
structure. In off-site application of intumescent coating,steel
members together with the welded connection componentsare sprayed
with intumescent coating and then assembled on site.Off-site
application of intumescent coating is gaining popularityowing to
better quality control and possible saving in constructiontime.
Scope exists to improve the efficiency of both on- and off-site
application of intumescent coating fire protection
withoutcompromising fire-resistance performance of the joints. In
thisresearch, fire tests on joints with different schemes of
intumes-cent coating fire protection were conducted to investigate
theeffects of the following three possible methods of reducing
fireprotection to joints:(1) Full protection of joint components
and connected membersexcept bolts.(2) Protecting a short segment of
the connected beams. This isrelevant to joints in steel-framed
structures where theconnected beams are unprotected but the joint
and thecolumn are protected.(3) Protecting only the columns. This
is relevant to joints in steel-framed structures where only the
columns require fireprotection due to their critical importance to
maintain globalstructural stability in fire.Conducting fire tests
on loaded specimens is extremely timeconsuming and expensive.
Therefore, the fire tests reported in thispaper were on unloaded
specimens. Nevertheless, since this paperis mainly concerned with
the relative temperature performance ofjoints with partial fire
protection compared to those with total fireprotection, it is
expected that the main conclusions of the paperwould be the same
regardless whether or not the specimens areloaded. In total, 10
steel joint specimens with different schemes offire protection were
fire tested in the gas furnace of the Universityof Manchester. In
an earlier part of this study [17], 4 steel jointspecimens without
any fire protection were also fire tested in thesame furnace. This
paper will mainly present the experimentalobservations and results
of the protected joints, but will also makereference to the
unprotected fire tests. A follow on paper willpresent the results
of analytical and numerical studies and thedevelopment of a design
method for calculating temperatures indifferent connection
components for implementation in subse-quent structural analysis of
joints using either finite-elementmethod or the component-based
joint characterisation method[18]. In addition, temperature
calculation is only part of theprocess of quantifying structural
fire resistance. Follow-onresearch studies are being carried out to
investigate the effectsof partial fire protection on structural
performance of joints.2. Research significance
Fire attack represents a significant risk to
steel-framedstructures. Until very recently, joint behaviour in
fire has receivedlittle attention from researchers, even though
joint behaviour hasa critical influence on controlling progressive
collapse of struc-tures under fire attack. Accurate prediction of
temperatures injoints represents the first step towards thorough
understandingand rational design of joints in steel-framed
structures in fire.From a practical point of view, efficient
application of intumescentfire protection can help to maintain the
advantages of steel-framed structures. This paper will report
results of temperaturesin partially protected steel joints in fire,
which will have practicalimplications on how cumbersome fire
protection to somecomponents of joints may be eliminated without
compromisingsafety of the structure.3. Test specimens and
set-up
3.1. Description of test specimens
A series of fire tests were conducted on steelconcrete
compositejoint assemblies with four types of joints; web cleat, fin
plate, flushendplate and flexible endplate, see Fig. 1. Four tests
were performedon joints without any fire protection and 10 tests on
joints withdifferent fire-protection schemes using intumescent
coating. Itshould be pointed out that to enable extensive
measurement oftemperatures and also to reduce the cost of fire
tests, the jointspecimens were not loaded. Each joint assembly
consisted of onecolumn, four beams (two bolted to the column
flanges via theaforementioned four different types of joints and
the other twobolted to the column web via fin plates) and a
concrete slab withprofiled steel sheeting and mesh reinforcement.
Fig. 1 shows the 3Dconfigurations of these joint assemblies. The
column in all the testspecimens was the same, being UC254254 89 and
with a lengthof 1000mm. The beam section in all the test specimens
was also thesame, being UB30516540. The length of the steel
beamsconnected to the column flanges was 605mm and the length ofthe
steel beams connected to the column web was 485mm. Table 1gives
detailed dimensions of joints for different test
specimens,including the four unprotected specimens identified as
USP1USP4and 10 protected specimens identified as SP1SP10. The
dimensionof the concrete slab was 10001000mm with an overall depth
of130mm. All columns, beams and joint components were in gradeS275
and grade 8.8 bolts of 20mm in diameter were used. To
enableconnections of different dimensions to be directly compared,
someof the tests used connectors of different dimensions in the
same testto join the two steel beams connected to the flanges of
the column.
3.2. Fire-protection schemes
As previously mentioned, the fire tests were designed
toinvestigate the effects of different fire-protection schemes on
thetemperature development within joint assemblies, focusing onthe
following three principal aspects: (1) not protecting bolts;(2)
protecting a small length of the beam near the joint and
-
ARTICLE IN PRESS
Fig. 1. Joint types and configurations: (a) web cleat joint
assembly; (b) fin platejoint assembly; (c) flush endplate joint
assembly and (d) flexible endplate joint
assembly.
Table 1Detailed dimensions of joint components
Specimen ID Join type to column flange Joint component to
one
column flange
USP1 Flush endplate
USP2 Flexible endplate
USP3 Fin plate 20010010
USP4 Web cleats 1509010 (depth200SP1 Web cleats
SP2 Web cleats
SP3 Web cleat
SP4 Fin plate 20015010SP5 Fin plate
SP6 Fin plate
SP7 Flush endplate
SP8 Flush endplate
SP9 Flexible endplate
SP10 Flexible endplate
X.H. Dai et al. / Fire Safety Journal 44 (2009) 376386378(3)
protecting the column only. Therefore, to achieve this aim,
thefollowing four fire-protection options were adopted to
theconnected beams:(1)32
2
)
32
2
Full protection including the bolts (to be referred to as FP+B):
theentire beam length (including the unconnected end cross-sections
of the beams) and the beam connectors including thebolts.(2) Full
protection but not including bolts (to be referred to as FP_B):as
in (1) but with the bolts unprotected.(3) Partial protection
including bolts (to be referred to as P300+B orP400+B): a segment
of 300mm (for beams connected to thecolumn web) or 400mm (for beams
connected to the columnflanges) of the beams from the corresponding
connection endswas protected, including all the beam connectors and
bolts.(4) Partial protection not including bolts (to be referred to
as P300_Bor P400_B): as in (3) but not including the bolts.The
columns were fully protected but the column connectorshave two
fire-protection options: (1) full protection includingbolts (FP+B)
and (2) full protection not including bolts (FP_B).
Fig. 2 shows two fire-protected specimens before testing
andTable 2 summarises the main features of the 14 tests, including
the 4unprotected specimens identified as USP1USP4. In Table 2,
Beams 1and 2 refer to the beams connected to the column flanges
(605mmlong) and Beams 3 and 4 refer to the beams connected to the
columnweb (485mm long). As shown in Fig. 2(a), the short beams
(Beams 3and 4) were perpendicular to the span of the decking.
Hence, theywere connected to the steel decking via shear
connectors. The longerbeams (Beams 1 and 2) were not connected to
the steel decking.Since intumescent coating was applied after the
steel decking wasalready connected to Beams 3 and 4, the upper
surface of Beams 3and 4 (which were connected to the steel decking)
could not becoated. In contrast, all surfaces of Beams 1 and 2 were
coatedaccording to the prescribed fire-protection schemes.
Intumescent coating fire protection was applied by
theintumescent coating manufacturers own application team.
Thenominal intumescent coating thickness was specified to limitthe
steel temperature rise to 550 1C at 60min of the standard
fireexposure to BS 476. Dry film thickness (DFT) measurements
weretaken at a number of locations in each specimen prior to
firetesting. Table 3 gives the average DFTs for different
specimens. Itcan be seen that although the beams and columns used
the samesection sizes in all tests, the actual average coating
thicknesses forJoint component to the
other column flange
Fin plates welded
to column web
420010
20010010
00150820015010
909010 (depth200)
20010010
420010
001508
-
ARTICLE IN PRESS
Fig. 2. Examples of different fire-protection schemes: (a)
unprotected bolts; (b) protected bolts.
Table 2Summary of fire-protection schemes
Specimen ID Join type to column flange Column Beam 1 Beam 2 Beam
3 Beam 4
USP1 Flush endplate Unprotected
USP2 Flexible endplate
USP3 Fin plate
USP4 Web cleat
SP1 Web cleat FP+B FP+B P400+B P300+B P300+B
SP2 Web cleat FP_B FP_B P400_B P300_B P300_B
SP3 Web cleat FP_NB None None None None
SP4 Fin plate FP+B FP+B P400+B FP+B FP+B
SP5 Fin plate FP_B FP_B P400_B FP_B FP_B
SP6 Fin plate FP_NB None None None None
SP7 Flush endplate FP+B FP+B P400+B FP+B P300+B
SP8 Flush endplate FP_B FP_B FP_B FP_B FP_B
SP9 Flexible endplate FP+B FP+B P400+B FP+B P300+B
SP10 Flexible endplate FP_B FP_B P400_B FP_B P300_B
Table 3Average dry film thickness (DFT) for different test
specimens
Specimen ID Joint type to column flange Average coat thickness
(DFT) in mm
Column Beam 1 Beam 2 Beam 3 Beam 4
SP1 Web cleat 0.67 1.02 1.18 1.12 1.15
SP2 Web cleat 0.73 1.02 1.29 1.05 1.09
SP3 Web cleat 0.60 / / / /
SP4 Fin plate 0.75 1.36 1.35 1.24 1.16
SP5 Fin plate 0.77 1.08 1.16 1.3 1.19
SP6 Fin plate 0.62 / / / /
SP7 Flush endplate 0.95 1.29 1.25 1.19 1.15
SP8 Flush endplate 0.84 1.14 1.19 1.43 1.35
SP9 Flexible endplate 0.78 1.21 1.19 1.25 1.33
SP10 Flexible endplate 0.86 1.22 1.16 1.21 1.31
X.H. Dai et al. / Fire Safety Journal 44 (2009) 376386
379different specimens varied considerably, being 1.021.43mm
onbeams and 0.60.95mm on columns. Also DFT measurements(not
presented here) at the different locations of the samemember showed
large variations.3.3. Test set-up
Fire tests were carried out in a gas-fired furnace
(internaldimensions: 3500mm3000mm2000mm) in the fire labora-tory of
the University of Manchester. The interior faces of thefurnace were
lined with ceramic fibre materials of thickness200mm that
efficiently transferred heat to the specimen. Two gasburners and
two exhausts were connected to the furnace. Thefurnace temperatures
were recorded by six conventional beadthermocouples. To ensure that
the concrete slab surface wasexposed to the ambient air
environment, the test specimen wasrotated by 901 and hung inside
the furnace via three well-protected steel ropes, as shown in Fig.
3. Fire exposure wasaccording to a standard fire condition
[19].
-
ARTICLE IN PRESS
Fig. 3. Set-up of joint test specimen: (a) view from inside the
furnace; (b) view from outside the furnace.
00 10 20 30 40 50 60
200
400
600
800
1000
Time (minutes)
Tem
pera
ture
(C
)
Standard Fire Furnace testUSP1 USP2USP3 USP4SP1 SP2SP3 SP4SP5
SP6SP7 SP8SP9 SP10
USP2
SP5
Fig. 4. Average furnace temperatures calculated using
furnace-control thermo-couples.
600
800
1000
atur
e (C
)
Standard fire Furnace testUSP2 USP3
USP2
X.H. Dai et al. / Fire Safety Journal 44 (2009) 3763863804.
Furnace temperatures
In order to compare recorded temperatures in joints obtainedfrom
different specimens, it was assumed that the temperaturefield
inside furnace was uniform and identical for different tests.To
verify this assumption, Fig. 4 shows the average gastemperatures
calculated using readings from the six monitoringthermocouples in
the furnace. It can be seen that the averagetemperatures in
different tests were very similar and close to theISO 834 [19]
standard fire temperature except for tests onspecimens USP2 and
SP5, where the average gas temperatureswere slightly higher than
the intended standard fire temperaturedue to failure of one control
thermocouple. In each test, the gastemperature field around the
test specimen was also measured bythermocouples around the test
specimen. Fig. 5 shows the averagegas temperatures calculated using
the thermocouples fixedaround the test specimens. It can be seen
from Fig. 5 thatalthough there is some deviation from the intended
standard firetemperature curve, the difference is small (typically
less than50 1C), indicating that gas temperature in the fire-test
furnace wasalmost uniform. To further confirm this, a supplementary
fire testwas carried out, in which there was no test specimen
butmeasurements were taken for gas temperatures near where thetest
specimen would be placed. Fig. 6 shows the measured gastemperatures
at different locations. The differences were small,further
confirming uniformity of gas temperature around thetest specimen.00
10 20 30 40 50 60
200
400
Time (minutes)
Tem
per
USP4 SP1SP2 SP3SP4 SP5SP7 SP8SP9 SP10
Fig. 5. Average furnace temperatures calculated using
thermocouples around test5. Temperature distributions in connection
components withdifferent fire-protection schemes
A large amount of data was collected. As previously
mentioned,the focus of this paper is to analyse the effects of
three differentpartial fire-protection schemes. The results will be
presentedunder these three headings.specimens.5.1. Effects of
partial protecting beams
It is now well accepted that a significant number of beams
insteel-framed buildings may be left unprotected, through the use
offire engineering design methods such as tensile membrane actionin
composite floor slabs [2022], control of load ratio or fireseverity
[23]. Yet in the majority of buildings using unprotectedsteel
beams, the columns and joints would normally have to beprotected to
ensure stability of the whole structure and to preventprogressive
collapse in fire. Under this circumstance, it isnecessary to decide
a suitable length of the beams for fireprotection so that the cost
of fire protection of a short segment ofthe beams is minimal, yet
the columns and joints perform as iffully protected without
suffering high temperature rises due toheat conduction from the
unprotected beams away from thejoints. In the intumescent coating
industry, a length of 400mmhas been considered adequate, although
the basis of this practice
-
ARTICLE IN PRESS
X.H. Dai et al. / Fire Safety Journal 44 (2009) 376386 381is not
clear. To investigate the validity of this practice, a series
ofjoints, described in the previous sections, with partially
protectedbeams (400 and 300mm from the connection zone) were
firetested.
Figs. 7 and 8 compare typical temperature distributions in
theweb cleat connected to the column flange of test SP1
(represent-ing web cleats and end plates in the column flange zone)
and inthe bolts of the fin plates in test SP4 (representing web
cleats/finplates/bolts/beam web in the beam web zone). The
closeness oftemperature curves (typical difference being about 20
1C) betweenwith full beam fire protection and with partial beam
fire00 10 20 30 40 50 60
200
400
600
800
1000
Time (minutes)
Tem
pera
ture
(C) T1 T3T4 T5
T6 T7T8 T10T11 T12T14 T15T16 T17T18 T19T20 T21T22 T23T24 T25T26
T27T-Average T-Standard fire
Fig. 6. Temperature distribution inside the furnace from a
supplementary fire testwithout a joint specimen.
UB305X165X40
J-J View
4020
0
150
4060
6040
10
50 50 50
UC254X254X89
90
50 40
J
J
X66
Web cleat150x90x10mm
605
0
100
200
300
400
500
600
Time (m
Tem
pera
ture
(C)
TC66 (full beam
TC51 (partial b
0 10 20 30
Fig. 7. Comparison of temperatures in web cleats connected to
column flanges (Tprotection suggests that a length of 400mm of fire
protectionon the beams would be sufficient to achieve effective
full fireprotection to the connection components within 60min of
fireexposure. Fig. 9 compares typical temperatures in a fin
platewelded to the column web, where one beam was fully
protectedand the other was protected for a length of only 300mm
from thejoint end. It appears that temperatures associated with
thispartially protected beam were noticeably higher than
tempera-tures associated with the fully protected beam
(temperaturedifference 450 1C), indicating that it would not be
sufficient toachieve full protection to the joint if only 300mm of
the adjacentbeam was protected.5.2. Effects of partial protecting
bolts
In the general practice of steel structure construction,
steelcomponents are fabricated in shops and then transported to
sitefor erection. When using off-site intumescent coating,
thefabricated steel components (including connection components)are
usually applied with intumescent coating in the fabricationfactory.
When protection of the bolts is necessary, this is carriedout on
site after the steel frame is assembled using pre-coatedmembers and
connection components. Application of intumes-cent coating to bolts
on site may not be welcomed because of thetime necessary for site
operation as well as the difficulty toguarantee quality of on-site
application of intumescent coating tobolts. However, if the bolts
are left unprotected, they may beoverheated in fire and cause
structural failure. Therefore, if boltsare to be unprotected on
site, it is important that the effects of thisaction are fully
understood and considered in design. The effectsI-I View
90
50 40 4020
0
4060
6040
10
90
50 40
UC254X254X89
UB305X165X40
X52 Web cleat
90x90x10mm
I
I
X51
UncoatedZone
400605
inutes)
protection)
TC52 (partial beam protection: 400mm)
eam protection: 400mm)
40 50 60
est SP1): (a) locations of thermocouples (TC); (b)
temperaturetime curves.
-
ARTICLE IN PRESS
UB305X165X4040
505050
4060
6040
10UC254X254X89
X Fin plate(200x150x10mm)
40
50 50
4060
6040
10
Fin plate(200X
100X10mm)
UB305X165X40
UncoatedZone
400605
605
0
100
200
300
400
500
600
700
800
0 10 20 30 40 50 60Time (minutes)
Tem
pera
ture
(C)
TC17 (partial beam protection: 400mm)
TC24 (full beam protection)TC18 (partial beam protection:
400mm)
TC25 (full beam protection)
X18
X25
17X
24
Fig. 8. Comparison of temperatures in bolts in fin plates welded
to column flanges(Test SP4): (a) locations of thermocouples (TC);
(b) temperaturetime curves.
0
100
200
300
400
500
600
700
Time (minutes)
Tem
pera
ture
s (C
) TC16 (partial beam protection: 300mm)
TC11(full beam protection)
TC17 (partial beam protection: 300mm)
TC12 (full beam protection)
X16
X17Beam3 Beam4
Unc
oate
d Zo
ne
300
UB305X165X40
485
40
5050
4060
6040
UC254X254X89Fin plate (200x100x10mm)
X11
12X
485
4060
6040
0 10 20 30 40 50 60
Fig. 9. Comparison of temperatures in fin plates welded to
columnweb (Test SP7):(a) locations of thermocouples (TC); (b)
temperaturetime curves.
UB305X165X40
4020
0
150
4060
6040
10
50 50 50
UC254X254X89
J
J
X60
Web cleat150x90x10mm
X58
605
00 10 20 30 40 50 60
200
400
600
800
1000
Time (minutes)
Tem
pera
ture
(C)
TC60TC58 Test USP4, unprotected specimen
Test SP2, protection scheme: FP_B
Test SP1, protection scheme: FP+B
Fig. 10. Comparison of temperatures in the web cleat on the beam
web (TestsUSP4, SP1 and SP2): (a) locations of thermocouples (TC);
(b) temperaturetime
curves.
X.H. Dai et al. / Fire Safety Journal 44 (2009) 376386382of
unprotected bolts on joint and steel frame structural behaviourare
being investigated by the authors research group. This paperwill
focus on temperature developments.5.2.1. Temperature distributions
in connection components on the
beam web
Figs. 10 and 11 compare typical temperature developmentin web
cleats and associated bolts (representing connec-tion components on
the beam web) between the followingthree fire-protection options:
(1) unprotected (UP); (2) fullprotection of web cleats and bolts
(FP+B) and (3) full protec-tion of web cleats but no protection of
bolts (FP_B). From theresults in Figs. 10 and 11, it is possible to
make the followingobservations:(1) If all the connection components
were unprotected, theconnection temperatures were much higher than
in connec-tions with fire protection regardless of whether or not
thebolts were protected.(2) Fig. 10 shows that if the connection
components (web cleats)were protected, then not protecting the
bolts had a minorinfluence on temperature development in the
protectedconnection components (web cleats). In fact, the
differencein web cleat temperatures between the two different
tests(SP1: with fire protection to bolts; SP2: without fire
protectionto bolts) was no greater than the difference in web
cleattemperatures at different locations (TC58 and TC60) in thesame
test (SP1 or SP2). Results from tests on fin plateconnections were
very similar.(3) Comparing the bolt temperatures in Fig. 11, it is
clear that ifthe connection components (web cleats) were
protected,whether or not protecting the bolts had noticeable
effectson the bolt temperatures. As expected, not protecting the
bolts(Test SP2) generated higher bolt temperatures than
protectingthe bolts (Test SP1). Nevertheless, the difference in
bolttemperatures in these two tests (about 100 1C at 60min) is
-
ARTICLE IN PRESS
Tem
pera
ture
(C)
Fig. 1locat
X.H. Dai et al. / Fire Safety Journal 44 (2009) 376386 383much
less than the difference in bolt temperatures betweentests on
unprotected connection (USP4) and fully protectedconnection
including bolts (Test SP1), being about 450 1C at60min. Results
from tests on fin plate connections weresimilar. Fig. 12 shows the
unprotected bolts in a protected webcleat before and after the
test. The bolts were not covered bythe expanded intumescent char so
the conclusions of thesetests can be generally applied.5.2.2.
Temperature distributions in connection components on the
column flange
Figs. 1315 compare typical temperatures in the
connectioncomponents and in the bolts connected to column flanges.
Theresults in Figs. 1315 suggest identical trends as described in
thelast section. In fact, the results in Fig. 14 indicate
higherFig. 12. Unprotected bolts in web cleat before
UB305X165X40
4020
0
150
4060
6040
10
50 50 50
UC254X254X89
J
J
X
X
61
56
Web cleat150x90x10mm
605
00 10 20 30 40 50 60
200
400
600
800
1000
Time (minutes)
TC61TC56
Test USP4, unprotected specimen
Test SP2, protection scheme: FP-B
Test SP1, protection scheme: FP+B
1. Comparison of temperatures in the bolts in web cleat on the
beamweb: (a)ions of thermocouples (TC); (b) temperaturetime
curves.connection temperatures when with full fire protection than
withno fire protection on bolts. This is mainly caused by
inconsistentintumescent coating behaviour such as coating
thickness, but itdoes suggest that not protecting the bolts did not
have asignificant detrimental effect. Furthermore, Fig. 15 shows
that atthe early stage of the fire exposure, the temperature
increasesin the unprotected bolts (fire-protection scheme FP_B)
weremuch faster than in the protected bolts (fire-protection
schemeFP+B). However, the differences in bolt temperatures
almostdisappeared in the later stage of fire test. This is clearly
a result ofthe expanded intumescent coating char covering the
unprotectedbolts, as shown in Fig. 16(a). Unfortunately, it is not
alwayspossible to rely on the expanded intumescent char to
coverunprotected bolts. In most cases, the bolts were not coveredby
the expanded intumescent coating char as shown in Figs. 12and
16(b).and after 60min fire exposure (Test SP2).
00 10 20 30 40 50 60
200
400
600
800
1000
Time (minutes)
Tem
pera
ture
(C)
TC64TC63
Test USP4, unprotected specimen
Test SP2, protection scheme: FP_B
Test SP1, protection scheme: FP+B
UB305X165X40
J-J View
4020
0
150
4060
6040
10
50 50 50
UC254X254X89
90
50 40
J
J
64
63X
X
Web cleat150x90x10mm
605
Fig. 13. Comparison of temperatures in the web cleat on column
flange: (a)locations of thermocouples (TC); (b) temperaturetime
curves.
-
ARTICLE IN PRESS
00 10 20 30 40 50 60
200
400
600
800
1000
Time (minutes)
Tem
pera
ture
(C)
Test USP1, unprotected specimen
Test SP7, protection scheme: FP+B
Test SP8, protection scheme: FP_B
6070
134
60
10
10605 200
X
50
25
E-E ViewE
E
UB305X165X40
45
Fig. 14. Comparison of temperatures in the flush endplate: (a)
locations ofthermocouples (TC); (b) temperaturetime curves.
00 10 20 30 40 50 60
200
400
600
800
1000
Time (minutes)
Tem
pera
ture
(C)
Test USP1, unprotected specimen
Test SP8, protection scheme: FP_B
Test SP7, protection scheme: FP+B
6070
134
60
10
10605 200
X
50
E-E ViewE
E
UB305X165X40
48
Fig. 15. Comparison of temperatures in the bolts in flush
endplate connections: (a)locations of thermocouples (TC); (b)
temperaturetime curves.
Fig. 16. Unprotected bolts in protected endplates after 60min
fire exposure:(a) bolts being covered by char after fire test (Test
SP8); (b) bolts not covered by
char after fire test (Test SP10).
00 10 20 30 40 50 60
200
400
600
800
1000
Time (minutes)
Tem
pera
ture
(C)
Test USP4, unprotected specimen
Test SP3, protecting column only
Test SP1, total protection
UB305X165X40
4020
0
150
4060
6040
10
50 50 50
UC254X254X89
J
J
X60
Web cleat150x90x10mm
X58
605
Fig. 17. Temperatures in the web cleat on beam web: (a)
locations of thermo-couples (TC); (b) temperaturetime curves.
X.H. Dai et al. / Fire Safety Journal 44 (2009) 3763863845.3.
Effects of protecting columns only
Because columns are critical structural elements, they
gen-erally will require fire protection. In the case of
unprotectedbeams and joints, the following question arises: would
connectioncomponents benefit from fire protection to the adjacent
column?As shown in Table 2, two tests (SP3 and SP6) adopted the
columnonly fire-protection scheme for web cleat and fin plate
connec-tions. This section presents the effects of protecting
columns onlyon the temperature distribution in these two types of
connectioncomponents. Figs. 17 and 18 compare temperatures for the
webcleat connections between full protection (FP+B, Test
SP1),protecting column only (PC, Test SP3) and no fire protection
(UP,Test USP4). Fig. 17 shows that for the connection components
onthe beam web, protecting column only was of no benefit inreducing
the connection temperatures. On the other hand, Fig. 18indicates
that if the connection components were contacted to theprotected
column flange (Test SP3), the connection temperatureswere
noticeably lower than in the unprotected test specimen (TestUSP4).
However, since the connection temperatures in the columnonly
protection specimen were much higher than in the fullyprotected
case (Test SP1), it is probably prudent to ignore thebenefit of
reduction in connection component temperatures due
-
ARTICLE IN PRESS
00 10 20 30 40 50 60
200
400
600
800
1000
Time (minutes)
Tem
pera
ture
(C)
Test SP3, protecting column only
Test USP4, unprotected specimen
Test SP1, total protection
UB305X165X40
J-J View
4020
0
150
4060
6040
10
50 50 50
UC254X254X89
9050 40
J
J
64
63X
X
Web cleat150x90x10mm
605
Fig. 18. Temperatures in the web cleat connected to column
flange: (a) locations ofthermocouples (TC); (b) temperaturetime
curves.
00 10 20 30 40 50 60
200
400
600
800
1000
Time (minutes)
Tem
pera
ture
(C)
Test SP6, protecting column only
Test USP3, unprotected specimen
Test SP4, total protection
40
505050
4060
6040
10
20
19X
X
UB305X165X40
UC254X254X89
Fin plate (200x150x10mm)
Fig. 19. Temperatures in welds on column: (a) locations of
thermocouples (TC);(b) temperaturetime curves.
600
800
1000
ture
(C)
Test USP4, unprotected specimen
Test SP3, protecting column only
X.H. Dai et al. / Fire Safety Journal 44 (2009) 376386 385to
column protection only. Only Fig. 19 suggests that temperaturesin
the welds, on the protected column flange, were much lowerthan if
the joint assembly was completely unprotected (TestUSP3). Although
the weld temperatures were higher than in thefully protected test
specimen (Test SP4), the reduction in weldtemperature due to fire
protecting the column alone wassubstantial and may be usefully
exploited in joint design.
It is expected that for temperatures in the column in the
jointregion, if only the column was protected by intumescent
coatingand the connectionwas not protected, the fire
protectionwould notbe as effective as when the entire joint
assembly was protected,because the intumescent coating on the
connection side would notbe able to expand and the column flange
would receive conductedheat from the unprotected connection. Fig.
20 shows confirmation.Similar temperature trends were observed in
the test using fin plateconnections. For safe design of joints, the
column components ofthe joint should be assumed to be
unprotected.00 10 20 30 40 50 60
200
400
Time (minutes)
Tem
pera
Test SP1, fully protected joint assembly
Fig. 20. Temperatures in column in the joint region: (a)
locations of thermo-couples (TC); (b) temperaturetime curves.6.
Conclusions
This paper describes fire experiments on unloaded steelcon-crete
composite joints with four types of connection: web cleat,fin
plate, flush endplate and flexible endplate. Results arepresented
for representative temperature distributions in connec-tion
components to demonstrate the effects of different fireprotection
schemes, including protecting only a short segment ofthe connected
beams, not protecting bolts in a protected joint andprotecting the
column only. Based on comparisons and analyses oftemperature
distributions in various connection components, thefollowing main
conclusions may be drawn:(1) Protecting a segment of the connected
beams by about400mm from the joint appeared to be sufficient to
achievefull protection for the joint. As far as the joint is
concerned, theeffect is similar in protecting the joint with steel
of 400mm inlength. Thus, this conclusion indicates that the 400mm
steel
-
ARTICLE IN PRESS
X.H. Dai et al. / Fire Safety Journal 44 (2009)
376386386protection has similar thermal resistance (thickness
dividedby thermal conductivity) as the applied intumescent
coating.Further numerical investigations are being carried out
toassess whether this 400mm rule would be applicable underother
conditions, e.g. different coating thickness.(2) In a protected
joint with unprotected bolts, the unprotectedbolt temperatures were
higher than those with full boltprotection, but the bolt
temperatures were still much lowerthan those in completely
unprotected joints.(3) In a protected joint, leaving the bolts
unprotected had littleinfluence on temperatures in other connection
components,i.e. end plates/fin plates/web cleats/beam/column.(4)
Compared to unprotected joints, protecting the column onlyhad
little benefit in reducing the temperatures in connectioncomponents
except for the welds on the column.(5) Although protecting the
column only did reduce columntemperatures in the joint region
compared to a totallyunprotected joint, the column temperatures in
the jointregion were substantially higher than those in a
fullyprotected joint assembly due to the heat conducted from
theunprotected connection components and beams. It would beprudent
to assume the column is unprotected when calculat-ing column
temperatures in the joint region.Due to limitations in sources and
time, only representativeconnections were fire tested in this
research. Nevertheless, it isexpected that the above conclusions
would hold for connectionsof other practical dimensions. Further
research studies are beingconducted to develop a method for
calculation of temperatures indifferent types of joints with
different fire-protection schemes andalso to understand the
implications on structural behaviour byadopting different
fire-protection schemes.Acknowledgements
This research is funded by a research grant from the
UKsEngineering and Physical Science Research Council
(EP/C003004/1). The authors would like to thank Mr. Jim Gorst and
Mr. Jim Geefor assistance with the fire tests.
References
[1] Federal Emergency Management Authority (FEMA), World Trade
CenterBuilding Performance Study, FEMA, USA, 2002.
[2] Institution of Structural Engineers (ISE), Safety in Tall
Buildings and Buildingsof Large Occupancy, Institution of
Structural Engineers, London, 2002.
[3] British Standards Institution (BSI, 2003), BS 5950 (2003),
Structural Use ofSteelwork in Buildings, Part 8: Code of Practice
for Fire Resistant Design,British Standards Institution, London.[4]
K.S. Al-Jabri, I.W. Burgess, R.J. Plank, Spring-stiffness model for
flexibleend-plate bare-steel joints in fire, J. Constr. Steel Res.
61 (11) (2005)16721691.
[5] F.M. Block, I.W. Burgess, J.B. Davison, R.J. Plank, The
development of acomponent-based connection element for endplate
connections in fire, in:P. Vila Real, J.M. Franssen, N. Lopes
(Eds.), SIF06Fourth InternationalWorkshop on Structures in Fire,
Aveiro, Portugal, vol. 1, 2006, pp. 345356.
[6] I.W. Burgess, Connection modelling in fire, in: F. Wald, F.
Mazzolani, M.Byfield, D. Dubina, M. Faber (Eds.), Proceedings of
the EU COST C26 ActionWorkshop in Prague, 2007, pp. 2534.
[7] J. Ding, Y.C. Wang, Experimental study of structural fire
behaviour of steelbeam to concrete filled tubular column assemblies
with different types ofjoints, Eng. Struct. 29 (12) (2007)
34853502.
[8] L.S. da Silva, A. Santiago, P. Vila Real, A component model
for the behaviour ofsteel joints at elevated temperatures, J.
Constr. Steel Res. 57 (11) (2001)11691195.
[9] S. Spyrou, J.B. Davison, I.W. Burgess, R.J. Plank,
Experimental and analyticalinvestigation of the compression zone
component within a steel joint atelevated temperatures, J. Constr.
Steel Res. 60 (6) (2004) 841865.
[10] S. Spyrou, J.B. Davison, I.W. Burgess, R.J. Plank,
Experimental and analyticalinvestigation of the tension zone
component within a steel joint at elevatedtemperatures, J. Constr.
Steel Res. 60 (6) (2004) 867896.
[11] Y.C. Wang, J. Ding, Experimental behaviour of steel joints
to concrete filledsteel tubular column in fire, in: The Fourth
International Symposium on SteelStructures, 1618 November 2006,
Seoul, Korea, 2006.
[12] H.X. Yu, I.W. Burgess, J.B. Davison, R.J. Plank,
Experimental investigation of thebehaviour of fin plate connections
in fire, in: Wang, Choi (Eds.), Proceedings ofthe Third
International Conference on Steel and Composite
Structures(ICSCS07), Taylor & Francis Group, London, 2007.
[13] J. Ding, Y.C. Wang, Temperatures in unprotected joints
between steel beamsand concrete-filled tubular columns in fire,
Fire Saf. J., 2008, in press,doi:10.1016/j.firesaf.2008.02.004.
[14] F. Wald, et al., The temperature distribution in a
full-scale steel framedbuilding subject to a natural fire, in: The
Second International Conference onSteel and Composite Structures
(ICSCS04), Seoul, Korea, September 24, 2004.
[15] F. Wald, L. Simoes da Silva, B.B. Moore, et al.,
Experimental behaviour of asteel structure under natural fire, Fire
Saf. J. 41 (7) (2006) 509522.
[16] F. Wald, M. Pultar, J. Chlouba, Fire tests on industrial
building. Steel andComposite Structures, in: Y.C. Wang, C.K. Choi
(Eds.), Proceedings of the ThirdInternational Conference on Steel
and Composite Structures (ICSCS07), Taylor& Francis Group,
London, 2007.
[17] X.H. Dai, Y.C. Wang, C.G. Bailey, Temperature distribution
in unprotected steelconnections in fire, in: Y.C. Wang, C.K. Choi
(Eds.), Proceedings of the ThirdInternational Conference on Steel
and Composite Structures (ICSCS07), Taylor& Francis Group,
London, 2007.
[18] Committee of European Normalisation (CEN, 2005), BS EN
1993-1-8: 2005,Eurocode 3: Design of steel structuresPart 18:
Design of joints, BritishStandards Institution, London.
[19] ISO 834, Fire Resistance Tests, Elements of Building
Construction, Interna-tional Organization for Standardization,
Geneva, 1975.
[20] C.G. Bailey, D.B. Moore, Structural behaviour of steel
frames with compo-site floorslabs subject to fire: part 1: theory,
Struct. Eng. 77 (11) (2000)1927.
[21] C.G. Bailey, D.B. Moore, Structural behaviour of steel
frames with compositefloorslabs subject to fire: part 2: design,
Struct. Eng. 77 (11) (2000) 2833.
[22] G.M. Newman, J.T. Robinson, C.G. Bailey, Fire Safe Design;
A New Approach toMulti-storey Steel-Framed Buildings (second
edition), SCI Publication P288,The Steel Construction Institute,
Ascot, 2006.
[23] Y.C. Wang, Steel and Composite Structures, Behaviour and
Design for fireSafety, Spon Press, London, 2002.
dx.doi.org/10.1016/j.firesaf.2008.02.004
Effects of partial fire protection on temperature developments
in steel joints protected by intumescent
coatingIntroductionResearch significanceTest specimens and
set-upDescription of test specimensFire-protection schemesTest
set-up
Furnace temperaturesTemperature distributions in connection
components with different fire-protection schemesEffects of partial
protecting beamsEffects of partial protecting boltsTemperature
distributions in connection components on the beam webTemperature
distributions in connection components on the column flange
Effects of protecting columns only
ConclusionsAcknowledgementsReferences