experimental and numerical modelling of cellular beams...

Volume 2 · Number 4 · 2011

289

Experimental and NumericalModelling of Cellular Beams with

Circular and Elongated WebOpenings at Elevated TemperaturesEl-Hadi Naili*, Ali Nadjai, Sanghoon Han, Faris Ali and Sengkwan Choi

University of Ulster, Jordanstown Compus, School of Built Environment Shore Road, Co-Antrim, Northern Ireland, BT37 0QB, UK

e-mail*: [email protected]

ABSTRACTThis paper describes an experimental and numerical study at elevated temperatures onthe behaviour of full scale composite floor cellular steel beams with circular andelongated web openings. A total of three specimens, comprising three different steelgeometries and loading conditions were tested at elevated temperatures. Finite elementmodels were established with both material and geometrical non-linearity to comparewith the experimental results. This paper also demonstrates the capability of a developedsimple design approach in comparison with numerical modelling, experimental tests andexisting design software used by the Steel Construction Institute (SCI).

Keywords: Cellular steel beams, Circular and elongated web openings, Fire testing, FEmodelling

1. INTRODUCTIONCellular beams (CBs) are currently being widely used in multi-storey buildings where, as well asreducing the total weight of the steelwork, they decrease the depth of floors by accommodating pipes,conduits and ducting. They are also used in commercial and industrial buildings, warehouses, andportal frames. CBs produced by modern automated fabrication processes can be competitive for theconstruction of both floor and roof systems. Their widespread use as structural members has promptedseveral investigations into their structural behaviour.

Lawson [1] presented a design method for simply-supported composite beams with rectangularopenings in the web at ambient temperature. The design method is based on plastic analysis of thecross-sections, considering the moment transfer by Vierendeel action across openings. Chung et al. [2]have investigated the Vierendeel mechanism in steel beams with circular web openings based onanalytical and numerical studies. Chung and Lawson [3] have presented a simplified design method inthe format of application rules to Eurocode-4 [4] for the design of simply-supported composite beamswith large web openings. Liu and Chung [5] have carried out a non-linear finite element analysis (FEA)investigation on steel beams with various shapes and sizes of web openings. Details of this designmethod are fully presented in a complementary paper [6]. Bailey [7] has investigated the temperaturesexperienced by the web-posts on cellular beams by carrying out preliminary indicative fire tests onunprotected and protected cellular and solid-web steel beams. Bitar et al. [8] proposed a model for webpost resistance based on experimental studies and numerical investigations which covers symmetricaland unsymmetrical sections. Lawson et al. [9] developed a design method for composite asymmetriccellular beams, which is not fully covered by existing design rules. Nadjai et al. [10, 11] carried outfull-scale fire tests and numerical studies at both ambient and elevated temperatures. Four specimens,comprising two different steel geometries and loading conditions were tested at elevated temperatures.Vassart et al. [12] have conducted an extensive parametric study based on the tests results of four full-scale fire tests that have been conducted on composite cellular beams with circular and/or

elongated web openings [13]. They have developed an analytical model that can be used for theprediction of the critical temperature of cellular beams. A full scale fire test has been performed recentlyon a composite floor for analysing the possibility of tensile membrane action to develop when theunprotected steel beams in the central part of the floor are made of cellular beams [14].

This paper presents the experimental and numerical studies on cellular beams with circular andelongated web openings at elevated temperatures which have the potential to provide essential data inseveral areas currently lacking systematic research. The target of this study is to investigate andunderstand the performance and failure mechanisms under a standard heating regime of cellular beamswith temperature distribution through the specimens.

2. TEST PROGRAMMEThe tests were carried out on three full-scale composite cellular steel beams of 4.5 m span length. Theywere fabricated from standard hot-rolled steel sections, subjected to one or two point loading, usingthree different geometries.

2.1. Specimens DetailsThe following types of beams have been tested:

a) Test 1A: An asymmetrical composite cellular beam was produced on the basis of UB 356 × 171 ×57 as a top tee section and of UB 610 × 305 × 179 as a bottom tee section having a finished depthof 555 × 171/305 ACB × 118 kg/m (see Figure 1). The diameter of the cells was 375 mm at600 mm centres.

b) Test 2A: A symmetrical composite cellular beam was produced on the basis of UB 457 × 191 × 74,having a finished depth of 550 × 191 CB 74 kg/m (see Figure 2). The diameter of the cells was335 mm at 600 mm centres.

c) Test 3A: An asymmetrical composite cellular beam was produced on the basis of UB 356 × 171 × 57as a top tee section and of UB457 × 191 × 74 as a bottom tee section having a finished depth of 555 ×171/191 ACB × 65.5 kg/m (see Figure 3). The diameter of the cells was 375 mm at 600 mm centres.

The steel grade of the beams was given as S355. All the tests used a 150 mm thick × 1100 mm wideconcrete slab, using normal-weight concrete (Grade 35 N/mm2). The slab reinforcement consisted ofwelded wire mesh reinforcement A142 having yield strength of 460 N/mm2. The interaction betweenthe slab and beam was ensured in all specimens by the use of shear connectors of 19 mm diameter studsat height 95 mm. They have been equally distributed in one row with a distance of 150 mm over thebeam length. The degrees of shear connection calculated are 0.53, 0.73 and 0.73 for the specimens 1Aand 2A and 3A, respectively. Holorib sheets HR 51/150 with a thickness of 0.9 mm have been used assheeting. The measured yield stress from a tensile yield stress from a tensile test was fy = 327 N/mm2.

290 Experimental and Numerical Modelling of Cellular Beams with Circular andElongated Web Openings at Elevated Temperatures

Journal of Structural Fire Engineering

Applied load

Beam detail:

Top tee:Tee depth = 255.0 mmWeb thickness = 8.1 mmFlange width = 172.2 mmFlange thickness = 13 mmBottom tee:Tee depth = 300.0 mmWeb thickness = 14.1 mmFlange width = 307.1 mmFlange thickness = 23.6 mm

4500 mm span375 DIA @ 600 spacing − 150 mm slab thickness555 Deep ACB S355 − 356 × 171 × 57 Top, 610 × 305 × 179 bot30 No 19 DIA stud × 95 long @ 150 C/C = 4350

Elastic moment capacity of cellular steel beam, Ms = 487.22 kNmPlastic moment capacity of cellular steel beam, Mc = 718.60 kNm

Figure 1. Details of the asymmetrical composite beam 1A.

El-Hadi Naili, Ali Nadjai, Sanghoon Han, Faris Ali and Sengkwan Choi 291


Applied load

Beam detail:

Tee depth = 550.0 mmWeb thickness = 9.0 mmFlange width = 190.4 mmFlange thickness = 14.5 mm

4500 mm span335 DIA @ 600 spacing − 150 mm slab thickness550 Deep CB S355 − 457 × 191 × 7430 No 19 DIA stud × 95 long @ 150 C/C = 4350


Figure 2. Details of the symmetrical composite beam 2A.

Applied load Beam detail:

Top tee:Tee depth = 255.0 mmWeb thickness = 8.1 mmFlange width = 172.2 mmFlange thickness = 13 mm

Bottom tee:Tee depth = 300.0 mmWeb thickness = 9.0 mmFlange width = 190.4 mmFlange thickness = 14.5 mm

4500 mm span375 DIA @ 600 spacing − 150 mm slab thickness555 Deep CB S355 − 356 × 171 × 57 top, 457 × 191 × 74 bot30 No 19 DIA stud × 95 long @ 150 C/C = 4350


Figure 3. Details of the asymmetrical composite beam 3A.

Concrete compressive strength was measured at different stages of time: after 2 weeks and 28 daysusing a compressive strength calibrated machine at the University of Ulster. During the testing days,the concrete compressive strength was found approximately equal to 35 N/mm2. The geometry data ofthe beams tested are given in Table I.

Table I. Geometry data

Beam 1A Beam 2A Beam 3ASpan (mm) 4500 4500 4500Top tee width (mm) 172.2 190.4 172.2Top tee depth (mm) 255 275 255Bottom tee width (mm) 307.1 190.4 190.4Bottom tee depth (mm) 300 275 300Overall depth (mm) 555 550 555Number of circular cells 6 3 7Number of elongated cells 1 2 0Number of cells with infill 0 1 1Number of cells with semi infill 2 0 0Overall number of cells 7 4 6Cell diameter (mm) 375 335 375Cell spacing (mm) 600 600 600

2.2. Mechanical LoadingTwo beams were tested under one point loading and one beam was tested under two-points loading;both ends of the beams were simply-supported. The beams were designed at ambient temperature inorder to determinate the failure load. The applied loads during the fire tests were considered equal tothe load ratio of 30% that depends on the standard ISO curve used during the test. The load ratiorepresents 30% of the failure load found from the pre-design at cold condition and by taking intoaccount the loading applied during the previous tests conducted at Ulster University as reference [11].During the fire tests, the composite cellular beams 1A, 2A and 3A were subjected to 200 kN, 150 kNand 150 kN constant loading, respectively.

2.2.1. Initial Design based on the SCI documentationThe specimens were pre-designed using software which is based on the SCI documentation [15] for thedesign of composite cellular beams and gives a linear elastic response in order to evaluate the failureload at cold conditions. Vierendeel bending, web post buckling and horizontal shear are considered asthe main failure modes of the three specimens. Table II gives the unity factor which is the degree ofutilisation of the beams in the failure modes and the failure loads. From the results, the Vierendeelmechanism occurs at the initial stage of failure before buckling of the web post and the beams failbecause of Vierendeel bending. The failure loads evaluated by the software are based on the first stageof failure only during the elastic phase of response.

2.2.2. Finite Element modelling at ambient temperatureGeometrically nonlinear finite element analysis have been carried out using the the commercial FEApackage Diana with non-linear material properties in order to simulate the complete behavior ofcellular beams at cold condition and validate how well it can predict the failure load by comparingpredictions with the previous cold tests [11] which were used as reference. Shell elements with theability to handle large strains, large deformations, and plasticity were used to model the cellular steelbeams. Solid-brick elements were used in the analysis to improve the rate of convergence,incorporating a smeared crack approach for the concrete, to model the composite slab. For only theevaluation of failure load at ambient temperature, full interaction between the cellular steel beam andthe concrete slab was assumed in the model and the steel decking shape and the shear studs were notconsidered in the model. This assumption is also justified from test observations [10] whichconfirmed that no stud failure occurred before web-post buckling of the beam.

Figures 4 & 5 show the predicted load-deflection- relationships and the ultimate failure loads of thebeams. A linear elastic response can be seen in the load deflection curves at the initial loading stage.The first yield occurs at a load level of approximately 382 kN, 245 kN and 432 kN for beams 1A, 2Aand 3A, respectively (approximately 58%, 54% and 76% of the ultimate failure load determined byFEA). By introducing a web imperfection into the model, the sections failed at a smaller load and thefailure mechanism was closer to a Vierendeel mechanism with web post buckling, as expected. Themaximum deflection just before failure was between 10 mm and 20 mm, which should not exceed avalue of span/200 [16].

2.2.3. Simple approachA simple method was proposed by Chung et al. [2] to provide a rule for practical design of cellular steelbeams. The failure load can be estimated by considering that the beam alone contributes to the strength



Table II. Software results

Vierendeel Web post Horizontal Failure load bending buckling shear (kN)

Beam 1A 90% 51% 102% 382Beam 2A 102% 35% 44% 245Beam 3A 111% 73% 88% 432

of the specimen and the beam and slab act independently of each other, although it may estimate aconservative value for the load carrying capacity. In this case, the slab is considered as a uniformdistributed loading along the length of the beam. The method uses an empirical shear momentinteraction formula as follows:

(1)

where, Vsd, applied shear force, Vo,Rd, shear capacity of cellular section, Msd, applied moment, Mo,Rd,moment capacity of cellular section. This formula can be represented to give a simple empirical designrule to estimate the moment capacity of a cellular section, Mvo,Rd, under a global shear force, Vsd,against the Vierendeel mechanism as follows:

V

V

M

Msd

o Rd

sd

o Rd,

.

,

.

+

≤

2 5 2 5

1



Deflection (mm)

Deflection (mm)

0 20 40 60 80 100 120

Load

app

lied

(kN

)Lo

ad a

pplie

d (k

N)

0

100

200

300

400

500

600

700

800

0

100

200

300

400

500

600

700

800

0 20 40 60 80 100 120

(a) Deflected shape of beam 1A

(b) Load versus deflection of beam 1A

(c) Deflected shape of beam 2A

(d) Load versus deflection of beam 2A

CB - with slab (no imperfection)CB - with slab (imperfection)Failure load (software based on SCI - 382 kN)

CB - with slab (no imperfection)CB - with slab (imperfection)Failure load (software based on SCI - 245 kN)

Figure 4. Loading versus Deflection and failure loads of composite cellular Beam 1Aand 2A.

(2)

The failure loads found by this method and the properties of cellular beams are given in Table IIIbelow.

2.2.4. Results comparisonTable IV shows that both the FEA and the simple approach produce similar outcomes regardless of thedifferent loading cases for beams 1A and 2A. The values obtained by the software based on the SCIapproach, which gives a linear elastic response, are conservative when compared against the failureloads obtained by the FEA and the simple approach because the failure loads evaluated by the SCIsoftware are based on the first stage of failure only in the elastic phase of the response. The failure loadof Beam 2A obtained by the SCI software is very conservative because of the Vierendeel mechanismthat occurs at the initial stage of failure in Beam 2A which includes two elongated openings comparedto the other two beams.

M MV

VVo Rd o RdSd

o Rd, ,

,

. .

.= −

1

2 5 0 4

≥≥ MSd



Table III. Properties summary and failure loads

Mo,Rd(kNm) Vo, Rd (kN) Mvo, Rd(kNm) FL (kN)Beam 1A 718.60 409.00 562.02 574Beam 2A 611.65 370.94 5217.75 450Beam 3A 572.34 299.00 461.53 396

Deflection (mm) 0 20 40 60 80 100 120

Load

app

lied

(kN

)

0

100

200

300

400

500

600

700

800

CB - with slab (no imperfection) CB - with slab (imperfection)Failure load (software based on SCI - 435kN)

(a) Deflected shape of beam 3A

(b) Load versus deflection of beam 3A

Figure 5. Loading versus Deflection and failure load of composite cellular Beam 3A.

Table IV. Comparison of the failure load results

Simple approach (kN) Software based on SCI (kN) Diana FEM (kN)Beam 1A 574 382 [640; 690]Beam 2A 450 245 [430; 480]Beam 3A 396 432 [540; 600]

3. TESTS RESULTS3.1. Tests DurationThe three fire tests were carried out under the ISO834 fire curve. Only the lower side of the slab andthe steel section were fire-exposed. Table V shows the duration of the fire tests and the failure times:

3.2. Temperature Distribution and DeflectionThe average temperature distribution along the steel profile of the three specimens is shown inFigures 7 & 8. The average temperature of top flange is the coldest part of the steel profile with asignificant thermal gradient due to the slab on the top of it, and only the bottom part is exposed to firecomparing to the other parts of the steel section that were fire-exposed on 3 sides. The maximumtemperature values were recorded in the web, reaching up to 830°C in Test 1A after 49 minutes and upto 795°C in Test 2A and Test 3A after 39 minutes. Figures 7 & 8 show the mid-span deflections recordedduring the three tests; the beam responds linearly due to the severe rise in temperature until about20 minutes in Test 1A, by which time the furnace temperature had risen to over 780°C and until about15 minutes in both Test 2A and Test 3A when the temperature was around 730°C. After these points thebeams started to lose strength and the rate of deflection began to gradually increase as the temperaturein the furnace reached over 860°C in Test 1A and around 800°C in both Test 2A and Test 3A.

The rate of deflections of Beam 2A and Beam 3A begin to gradually increase after this point due tothe deterioration of the beam properties until about 24 and 26 minutes respectively, when the beamsdeflection is recorded at furnace temperature around 800°C. Between 20 and 25 minutes, for both Beam2A and Beam 3A rate of deflection started to increase rapidly up until the point of failure at 39 minutes,by which time the beams had deflected by 249 mm and 253 mm at furnace temperature around 870°C,respectively. The deflection of the Beam 1A continued to rise more rapidly as the beam lost morestrength and stiffness up until the point of failure after 49 minutes at furnace temperatures around920°C, with the deflection recorded equal to 254 mm.



Table V. Duration of fire tests

Test 1A Test 2A Test 3AHeating Phase (min) 60 43 50Failure Time (min) 49 39 39

The figure shows the thermocouple locations on web-post 1 from left side. 66 thermocouples were used on the steel section (Test 1A)

12 Thermocouples were located on the slab and shear studs

10 mm

10 mm

140 mm

140 mm2 × 10 mm

a

1 2

e

b m i

jk

l

c nd

g P

o A

A

B

B

C

C

3 off 4 off2 off

50 mm

50 mm

5 mm

5 mm5 mm

5 mm5 mm 5 mm

h f

Figure 6. Typical thermocouple locations on the steel and slab (Test 1A).



Time (minutes)

0 10 20 30 40 50 60 70

Ave

rage

tem

pera

ture

(°C

)

0

200

400

600

800

1000

Bot flange Bot webTop webTop flange Mean furnace temperatureISO834 curve

Time (min)

0 10 20 30 40 50 60

Def

lect

ion

(mm

)

−300

−250

−200

−150

−100

−50

0

Mid-span deflection

(a) Time vs. temperature in the steel of test 1A (b) Time vs. deflection of beam 1A

Figure 7. (a) Time versus temperature in the steel and (b) deflection - Test 1A.

Bot flange Bot webTop webTop flange Mean furnace temperatureISO834 curve

Bot flange Bot webTop webTop flange Mean furnace temperatureISO834 curve Mid-span deflection

Mid-span deflection

Time (minutes)0 10 20 30 40 50

Ave

rage

tem

pera

ture

(°C

)A

vera

ge te

mpe

ratu

re (

°C)

0

100

200

300

400

500

600

700

800

900

1000

Time (minutes) Time (min)

0 10 20 30 40 50 600

200

400

600

800

1000

Time (min)0 10 20 30 40 50 60

Def

lect

ion

(mm

)D

efle

ctio

n (m

m)

−300

−250

−200

−150

−100

−50

0

0 10 20 30 40 50 60−300

−250

−200

−150

−100

−50

0

(a) Time vs. temperature in the steel of test 2A (b) Time vs. deflection of beam 2A

(c) Time vs. temperature in the steel of test 3A (d) Time vs. deflection of beam 3A

Figure 8. (a) Time versus temperature in the steel and (b) deflection - Test 2A and (c)time versus temperature in the steel and (d) deflection - Test 3A.

3.3. Failure ModeAll sections of the fire tests failed due to the fact that they lost strength and stiffness due to the risein temperature. Figure 9 below shows the failure mechanism observed in the tests. The buckling ofthe web posts begins to occur before the final point of failure as the steel beam temperatures exceed600°C, at which point the steel has less than half of its design strength and its Young’s modulus isreduced to about 20% of the room temperature value. When the furnace temperature is around750°C, the Young modulus decreases more rapidly than the steel strength limit which is the causethe failure modes. The main failure mode in Test 1A and Test 2A was the Vierendeel bendingassociated with the buckling of the web posts of the steel section. Web post buckling was the mainfailure mode in the Test 3A.

4. FINITE ELEMENT MODEL FOR FIRE CONDITIONSThe cellular steel beam sections and slab were modelled using solid-brick elements and heatingelements in order to add a temperature dependent mesh over top of the structural mesh.

Both the steel deck, as a bottom layer, and the reinforcing mesh, as a layer within the concrete, wereincluded. To simulate the tests as accurately as possible the beams were split into different areas.



Test 1A: Web post buckling and vierendeel bending failure mechanisms

Test 2A: Vierendeel bending associated with web post buckling failure mechanisms

Test 3A: Web post buckling failure mechanism

Figure 9. Deformed beams after fire tests.

Different time/temperature curves were introduced to the model according to the average thermocouplereading recorded in the tests for the bottom flange, bottom web, upper web, upper flange, bottom layerof steel decking and concrete slab. A smeared cracking model was used for concrete which ischaracterized by the use of combining tension softening, tension cut-off and shear retention to analyzea concrete structure under loading. Tension cut-off has one of two options to consider, either constantor linear performance under loading, as shown in Figure 10. An implicit analysis was conducted in twosteps, where the load was applied in the first step and the temperature was applied in the second step.



(a) Constant (b) Linear

Crack 1 Crack 2

ƒc

ƒc

ƒt ƒt

ƒtƒt

σ2 σ2

σ1 σ1

Figure 10. Criteria for tension cut-off in smeared cracking FE model.

Deflection (mm)0 50 100 150 200 250

0

10

20

30

40

50

Deflection (mm)

0 50 100 150 200 250

Tim

e (m

in)

0

10

20

30

40

50

60

Test dataFEM- DIANA

Test dataFEM- DIANA

(a) Deflection at mid-span versus time (b) Deformed shape and failure mechanism

(c) Deflection at mid-span versus time (d) Deformed shape and failure mechanism

Figure 11. Comparison of FEM models and test results (a & b) Beam 1A and (c & d)Beam 2A.



Deflection (mm)

0 50 100 150 200 250

Tim

e (m

in)

0

10

20

30

40

50

Test dataFEM - DIANA

(a) Deflection at mid-span versus time (b) Deformed shape and failure mechanism

Figure 12. Comparison of FEM model and test result - Beam 3A.

Figures 11 & 12 show the load-deflection curves compared between FE models and the experimentaltests and the output deformed shape of FEM analysis which can be compared with the actual failuremodes of the three specimens.

The approach of numerical modelling agrees reasonably well with the experimental fire test results.The failure mode that has taken place in Beam 1A is due to the web posts buckling and Vierendeelbending, as was seen in the fire test. During the initial stage of loading, the Vierendeel mechanism tendsto develop, starting due to the cellular geometry of specimen. Finally, the beam fails by flexural webpost buckling.

The axial stress distribution in the elements at the early stages of failure is shown in Figure 13. Fromthe results, the stress concentration around the openings can help to determinate the critical sections inthe three beams tested. For Beam 2A, the stress concentration in the bottom part of the web around theelongated opening near the circular opening is in tension to a greater extent than the other parts of theweb around openings and the stress concentration in the top part of the elongated openings is higherthan in the web post. From these observations and analysis, the failure mode of Test 2A has clearly beenthe Vierendeel mechanism, as expected associated with web post buckling.

5. CONCLUSIONThis paper describes an experimental and numerical study of the behaviour of three CBs with circularand elongated web openings at elevated temparatures. Flexural web post buckling was the main failuremode in Test 1A of an asymmetrical composite beam in the area of moment and shear interaction. Onlythe top part of the web resists the loading due to the high asymmetrical geometry of the steel beam ofcomposite section 1A. Similar failure occurred in Test 3A of an asymmetrical composite beam.

.156E9

.954E8

.35E8−.254E8−.858E8−.146E9−.207E9−.267E9−.327E9−.388E9

Figure 13. Axial stress distribution in elements (N/m2 ) - Beam 2A.

Veirendeel bending associated with the buckling of the web post were the main failure modes in Test2A of a symmetrical composite beam. The FEM modelling predicted well the structural behaviour andfire resistance of each of the test specimens. The models can be used for further parametric studiesusing long span beams in order to extend existing fire design rules. Based on the observations andanalysis of stress distributions and deflected shapes in the three tested beams with different geometries,the critical sections of CBs can be applied to evaluate the effective length for analytical modeling ofweb post failures.

REFERENCES[1] R.M. Lawson, Design for openings in the webs of composite beams, CIRIA/SCI publication P068

(1987).

[2] K.F. Chung, T.C.H. Liu, A.C.H. Ko, Investigation on Vierendeel mechanism in steel beams withcircular web openings, Journal of Constructional Steel Research. 57 (2001) 467–490.

[3] K.F. Chung, R.M. Lawson, Simplified design of composite beams with large web openings toEurocode 4, Journal of Constructional Steel Research. 57 (2001) 135–164.

[4] ENV 1994-1-1: Eurocode 4: design of composite steel and concrete structures. BSI. (1994).

[5] T.C.H. Liu, K.F. Chung, Steel beams with large web openings of various shapes and sizes: finiteelement investigation, Journal of Constructional Steel Research. 59 (2003) 1159–1176.

[6] K.F. Chung, C.H. Liu, A.C.H. Ko, Steel beams with large web openings of various shapes andsizes: an empirical design method using a generalised moment-shear interaction curve, Journal ofConstructional Steel Research. 59 (2003) 1177–1200.

[7] C. Bailey, Indicative fire tests to investigate the behaviour of cellular beams protected withintumescent coatings, Fire Saf. J. 39 (2004) 689–709.

[8] D. Bitar, T. Demarco, P.O. Martin, Steel and non composite cellular beams – Novel approach fordesign based on experimental studies and numerical investigations, brochure EUROSTEEL(2005).

[9] R.M. Lawson, J. Lim, S.J. Hicks, W.I. Simms, Design of composite asymmetric cellular beamsand beams with large web openings, Journal of Constructional Steel Research. 62 (2006)614–629.

[10] A. Nadjai, Performance of Cellular Composite Floor beams at ambient temperatures, FireSERT,Test report, (2005).

[11] A. Nadjai, O. Vassart, F. Ali, D. Talamona, A. Allam, M. Hawes, Performance of cellularcomposite floor beams at elevated temperatures, Fire Saf. J. 42 (2007) 489–497.

[12] O. Vassart, C.G. Bailey, G. Bihina, M. Hawes, A. Nadjai, C. Peigneux, W.I. Simms, J.M.Franssen, Parametrical study on the behaviour of steel and composite cellular beams under fireconditions, Proceedings of the Sixth International Conference on Structures in Fire (SiF’10)(2010).

[13] G. Bihina, B. Zhao, O. Vassart, Cellular composite beams at elevated temperatures, Applicationof Structural Fire Engineering, Prague, Czech Republic (2009).

[14] O. Vassart, C.G. Bailey, M. Hawes, A. Nadjai, W.I. Simms, B. Zhao, T. Gernay, J.M. Franssen,Large-scale fire test of unprotected cellular beam acting in membrane action, Proceedings of theSixth International Conference on Structures in Fire (SiF’10) (2010).

[15] The Steel Construction Institute, Design of Composite and Non-Composite Cellular Beams(1994).

[16] BS 5950-1: 2000, Structural use of steelwork in building, Part 1: Code of practice for designrolled and welded sections.



experimental and numerical modelling of cellular beams...

Documents