WP4 –BENCHMARK STUDY - cvut.czpeople.fsv.cvut.cz/www/wald/fire/ifer/WP4/PP-Prague/03_Santiago.pdf · NUMERICAL BEHAVIOUR OF STEEL COLUMNS ... • ≠failure mode ( sway ) ... (EC3):

1

WP4 –BENCHMARK STUDY :NUMERICAL BEHAVIOUR OF STEEL COLUMNS

UNDER LOCALIZED FIRE LOADING

Cécile Haremza (WG2), Aldina Santiago (WG2), Fernanda Lopes (WG3), University of Coimbra

Jean-Marc Franssen (WG1), University of Liège

COST Action TU0904 Meeting

Prague, WG2, 18-19 April 2013

Institute for Sustainability and Innovation in Structural Engineering

2|WP4 - Column Benchmark Aldina Santiago

BENCHMARK STUDY : Numerical model of a steel column under a natural f ire

INTRODUCTION

Validation of the utilization of

ABAQUS and SAFIR for steel

structures under fire: thermal

and mechanical analyses

the model definition

Experimental results and the

numerical results obtained by

the software CEFICOSSagainst

FIVE STUDY CASES:

non-uniform temperature

axial restraint to beam

frame continuity

OBJECTIVE:

the reference case

2



BENCHMARK STUDY

� PAPER PUBLISHED BY Franssen, Cooke and Latham (1995): Fully loaded, 2D unprotected

steel frame under a natural fire loading

� COLUMN: 3530mm long, Grade 43A ( ≈S355), pin jointed at the

base

� BEAM: 4550mm long, Grade 43A

� BOLTED BEAM-TO-COLUMN JOINTS

� LATERAL AND SWAY INSTABILITIES PREVENTED

� CONCRETE BLOCKS between the

column flanges ( non-composite behaviour )

� CONCRETE SLAB ( non-composite behaviour )



MATERIAL PROPERTIES

� STEEL PROPERTIES

Eurocode 3, part 1.2 (EN 1993-1-2:2005)

At ambient temperature: fy = 408 MPa; E = 210 GPa

� CONCRETE PROPERTIES (Slab +

Blocks between column flanges):

Only plays an insulating role

0

50

100

150

200

250

300

350

400

450

0 0.05 0.1 0.15 0.2Strain

Stre

ss (M

Pa)

20 ºC100ºC200ºC300ºC400ºC500ºC600ºC700ºC800ºC900ºC1000ºC1100ºC1200ºC

3



THERMAL ANALYSIS

� ANALYSIS to determine the distribution of temperatures in each element

• Involves conduction and boundary radiation

• The concrete slab and the concrete blocks are modelled (thermal boundary conditions)

• 2-D SOLID FE elements

Mesh: ABAQUS SAFIR

Beam

Column

Y



0

100

200

300

400

500

600

700

800

900

0 5 10 15 20 25 30Time (min)

Tem

pera

ture

(ºC

)

IN flange_TestIN flange_CeficossIN flange_ABAQUSIN flange_SAFIROUT flange_TestOUT flange_CeficossOUT flange_ABAQUSOUT flange_SAFIRWeb_TestWeb_CeficossWeb_ABAQUSWeb_SAFIRAir

0

100

200

300

400

500

600

700

800

900

0 5 10 15 20 25 30

Time (min)

Tem

pera

ture

(ºC

)

Bottom flange_TestBottom flange_CeficossBottom flange_SAFIRBottom flange_ABAQUSTop flange_TestTop flange_CeficossTop flange_SAFIRTop flange_ABAQUSWeb_TestWeb_CeficossWeb_SAFIRWeb_ABAQUSAir

THERMAL RESULTS:

Column

Good correlation between CEFICOSS, SAFIR, ABAQUS an d

experimental resultsBeam

4



MECHANICAL ANALYSIS

� GENERAL MODELLING ASSUMPTIONS

• Static analysis

• 2D Beam elements

• Large-displacement effects and material nonlinearit ies

• Temperature dependent material properties (E, σσσσ−−−−ε, αε, αε, αε, α)

• Nonlinear temperature gradient

• Concrete (slab + blocks) not modelled

• Rigid beam-to-column connection

• Bi-linear spring (restraint – 2 nd steelwork)



MECHANICAL ANALYSIS

� MECHANICAL LOADING : (i) applied at room temp, (ii) maintained constant (fire)

� THERMAL GRADIENT THROUGH THE SECTION

• CEFICOSS, SAFIR: calculated temp. (thermal analysis) used by the str uctural part

• ABAQUS: 3 Dependent temperatures at three points

5



MECHANICAL ANALYSIS

� THERMAL GRADIENT ALONG THE BEAM SPAN (lower temp. at joint)

• CEFICOSS, SAFIR: sinusoidal function

• ABAQUS: temperature along each length part is constant

MidspanBeam-to-column connection

0.918 θθθθa 0.95 θθθθa 1.00 θθθθa0.974 θθθθa 0.99 θθθθa



� STUDY CASES

1. Reference case

2. Model definition

3. Axial restraint to beam

4. Frame continuity

5. Non-uniform distribution of temperature

MECHANICAL RESULTS

6



MECHANICAL RESULTS

� 1. REFERENCE CASE

Lateral deformation influenced by the concrete

blocks inserted between column flanges

0

500

1000

1500

2000

2500

3000

3500

4000

-500 0 500 1000 1500 2000 2500

Y (m

m)

X (mm)

Scale of displacements: 10/1

Initial frame

Experimental

SAFIR

ABAQUS

Frame deformation (t = 16 min)



MECHANICAL RESULTS

� 1. REFERENCE CASE

Due to the elongation of the beam �the column bows laterally up to the buckling around 19-20 min

-80

-70

-60

-50

-40

-30

-20

-10

0

0 5 10 15 20 25

Late

ral d

ispl

acem

ent

-m

id h

eigh

t (m

m)

Time (min.)

CEFICOSS

ABAQUS

SAFIR

16’

Rf time = 20’

Horizontal displacement at the column mid-height

Differences between programs: Slight ≠ temperature gradients in cross-sections

7



MECHANICAL RESULTS� 1. REFERENCE CASE - Beam mid-span vertical displacem ent

0

20

40

60

80

100

120

140

160

0 5 10 15 20 25

Ver

tical

dis

plac

emen

t -M

id s

pan

(mm

)

Time (min.)

Experimental Test

CEFICOSS

ABAQUS

SAFIR

0

20

40

60

80

100

120

140

0 5 10 15 20 25

Axi

al c

ompr

essi

on fo

rce

in b

eam

(kN

)

Time (min.)

CEFICOSS

ABAQUS

SAFIR

SAFIR: 125kN

CEFICOSS: 124kN

ABAQUS: 123kN



MECHANICAL RESULTS

� 2. INFLUENCE OF THE MODEL DEFINITION

Rf CEFICOSS Rf SAFIR Rf ABAQUS

1 - half of the frame 19’12’’ 20’04’’ 19’55’’

2 - complete frame 19’22’’ 17’55’’ 19’51’’

• Initial lateral imperfection : 0.8Hc/1000

• Fire resistance time:

• Good agreement between CEFICOSS and

ABAQUS

• ≠ SAFIR � difficulty in modelling the two

springs working together

8



0

20

40

60

80

100

120

140

0 5 10 15 20 25

Axi

al c

ompr

essi

on fo

rce

in b

eam

(kN

)

Fire duration (min.)

CEFICOSS_Restrained frame 1a

SAFIR_1a

ABAQUS_1a

CEFICOSS_Unrestrained frame 2a

SAFIR_2a

ABAQUS_2a

MECHANICAL RESULTS

� 3. INFLUENCE OF THE AXIAL RESTRAINT

ABAQUS: 45kN SAFIR: 43kN CEFICOSS: 43kN

• Axial compression force

• Fire resistance and Stability few affected

• GOOD CORRELATION



� 3. INFLUENCE OF THE AXIAL RESTRAINT TO BEAM – Comple te frame

MECHANICAL RESULTS

ABAQUS

Half-of-the frame

• ≠ failure mode (sway ) � Fire resistance reduced from 20 min. to 12-13 min.

tf ≈ 20’

SAFIRt f = 12’08’’

ABAQUSt f = 12’17’’

CEFICOSSt f = 13’45’’

• Good correlations between numerical programs

9



MECHANICAL RESULTS

� 4. INFLUENCE OF THE FRAME CONTINUITY



MECHANICAL RESULTS

� 4. INFLUENCE OF THE FRAME CONTINUITY – Beam on its o wn

Vertical displacement of the beam

0

20

40

60

80

100

120

140

160

180

200

0 5 10 15 20 25

Bea

m m

id-s

pan

vert

ical

dis

pl. (

mm

)

Time (min.)

EXP. TEST_Beam as part of the frame

SAFIR_Beam as a separate member

ABAQUS_Beam as a separate member

• Vertical displacement (no beneficial restraints from the column

• Reduction of the R f time (17’-18’ for SAFIR and ABAQUS)

• Small Discrepancies between FE programs SAFIR and ABAQUS

10



-60

-50

-40

-30

-20

-10

0

10

-5 0 5 10 15 20 25 30

Late

ral d

ispl

acem

ent -

mid

hei

ght (

mm

)

Fire duration (min.)

CEFICOSS_Column as a separate member

ABAQUS

SAFIR

CEFICOSS_Column as part of the frame

MECHANICAL RESULTS

� 4. INFLUENCE OF THE FRAME CONTINUITY – Column on its own

• All the programs: R f time > 30’ ( ≠ 20min.) Horizontal displacement of the column

• Column on its own: behaviour ≠

• Column remains stable - no collapse



MECHANICAL RESULTS

� 5. INFLUENCE OF THE NON-UNIFORM TEMPERATURE

0

100

200

300

400

500

600

700

800

900

0 5 10 15 20 25 30

Tem

pera

ture

(ºC

)

Time (min.)

Air temp.BEAM_Uniform temp._EC3BEAM_Aver. temp_CEFICOSSBEAM_Aver. temp_SAFIRBEAM_Aver. temp_ABAQUSCOLUMN_Uniform temp._EC3COL_Aver. temp_CEFICOSSCOL_Aver. temp_SAFIRCOL_Aver. temp_ABAQUS

• Gradient temp. (reference case)

• Uniform Temp through the cross-section: EN 1993-1-2:2005 - simplified method

• Average temperature

11



MECHANICAL RESULTS

� 5. INFLUENCE OF THE NON-UNIFORM TEMPERATURE - Column

-100

-90

-80

-70

-60

-50

-40

-30

-20

-10

0

0 5 10 15 20 25 30

Col

umn

late

ral d

ispl

acem

ent

-m

id-h

eigh

t (m

m)

Time (min.)

CEFICOSS_Non-uniform temp.

ABAQUS_Non-uniform temp.

ABAQUS_Uniform temp.

ABAQUS_Aver. temp

• Average temperature : leads to unsafe results (no failure is observed until the end of the analysis)

• Uniform temperature (EC3): leads to conservative results (premature failure)



CONCLUDING REMARKS

� SAFIR and ABAQUS performed well the heat transfer analysis

� SAFIR and ABAQUS showed a good ability to simulate steel structural behaviour under fire conditions using beam elements

� Some of the differences between the results of the 2 programs could be explained by:

• Temperature gradients in cross-sections approximate d in ABAQUS

• Difficulty to simulate the behaviour of springs wit h SAFIR (no dedicated finite element)

12

Documents

WP4 –BENCHMARK STUDY - cvut.czpeople.fsv.cvut.cz/www/wald/fire/ifer/WP4/PP-Prague/03_Santiago.pdf · NUMERICAL BEHAVIOUR OF STEEL COLUMNS ... • ≠failure mode ( sway ) ... (EC3):