NUMERICAL STUDY ON THE BEHAVIOR OF STEEL COLUMNS EMBEDDED ON BRICK WALLS SUBJECTED TO FIRE

International Symposium “Steel Structures:Culture & Sustainability 2010”

21-23 September 2010, Istanbul, Turkey

Paper No: 70

NUMERICAL STUDY ON THE BEHAVIOR OF STEEL COLUMNS

EMBEDDED ON BRICK WALLS SUBJECTED TO FIRE

Antonio M. CORREIA1,2

, João P. C. RODRIGUES2

and Valdir P. SILVA3

1 Superior School of Technology and Management of Oliveira do Hospital, Portugal

2 Faculty of Sciences and Technology of University of Coimbra, Portugal

3 Polytechnic School of the University of S. Paulo, Brazil.

ABSTRACT

Steel columns embedded on building brick walls are subjected to differential heating that lead to higher

thermal elongation on the exposed side of the cross-section. The strain field varies from the hot to the cold

side of the cross-section that induces a curvature in the columns towards the fire direction. Due to the

restraint to thermal elongation, provided by the surrounding structure, additional stresses will be generated in

the columns increasing their axial force.

A numerical study, using the finite element code ABAQUS, were carried out to study the behavior of steel

columns embedded on brick walls, subjected to fire and with restrained thermal elongation. Two column

cross-sections, two thicknesses of one-leaf brick walls per type of column cross-section and two orientations

of the buckling axis in relation to the wall surface, were tested. The results of these numerical simulations

were validated with the ones of fire resistance tests carried out at the University of Coimbra.

From this study it was concluded that the building walls on one hand had a detrimental effect leading in

certain cases to column thermal bowing and on the other hand a beneficial effect of column thermal

insulation.

Keywords: steel, column, fire, resistance, wall, buckling

INTRODUCTION

The behavior of steel columns is known to be strongly dominated by the loss of strength due to the

degradation of the mechanical properties with the increase of the temperature, namely the yield

strength and the Young’s modulus, and their interaction with the building structure in which they

are inserted.

EN1993-1-2(2005), presents a formulation for the calculation of the temperature evolution on steel

cross-sections totally engulfed in fire. The columns in real buildings are usually partially embedded

on walls that on one hand protect them from the excessive heating resulting from the fire and on

other hand induce high thermal gradients in the columns cross-section leading to differential strains

and thermal bowing.

The aim of this research is to study the influence of brick walls on the behavior of steel columns

subjected to fire. A great number of numerical simulations using the finite element program

ABAQUS were carried out. It was taken into account the material and geometric non-linearity of

the problem. They were tested H steel columns of different cross-sections, thickness of the brick

walls, and orientation of the buckling axis in relation to the wall surface and with and without

external loading.

The results of the numerical simulations were validated with the ones of some fire resistance tests

on steel columns embedded on walls, carried out at the Laboratory of Testing Materials and

Structures of the University of Coimbra. Several fire resistance tests were performed without

loading in order to assess the influence of the walls on the heating of the column while some were

performed with loading to study the structural behavior of the steel columns when embedded on

walls.

FIRE RESISTANCE TESTS

The test set-up consisted of a three-dimensional restraining frame (3) in which the specimen was

inserted (Figure 1). This frame that simulated the stiffness of the building structure to the column

subjected to fire was conceived in order to allow the consideration of different stiffness values.

However, the stiffness considered in these tests was 7kN/mm. The frame was built with beams and

columns HEA200, steel grade S355.

Figure 1. Test set-up for steel columns embedded on walls

The loading was applied by a hydraulic actuator (2) of 1MN controlled by a central unit. This

hydraulic actuator was mounted in a plane reaction frame (1). The loading was 70% of the design

value of the buckling load at ambient temperature, NbRd,20, calculated according to EN1993-1-

1(2005), and intended to simulate the serviceability load of the columns when in a real structure.

In both sides of the columns, brick walls (6) covered with 15mm of cement mortar in each surface

were built.

The heating was applied by a gas fire furnace (5) following the standard ISO 834 fire curve. The

thermal action was applied only on one side of the specimen.

Displacement transducers were used to measure the axial displacements in the columns during the

tests (4). They were placed at 3 different points, orthogonally arranged, in order to assess the

different planes of deformation of the columns, allowing with this, not only to measure the axial

displacements but also the rotations at their ends. The lateral displacements, in the direction

perpendicular to the wall surface, were also measured.

Temperatures inside and on the wall surface, at three different levels and deeps, and in the columns

were also measured. For that purpose type K thermocouples were used.

1

2

4

3

4

5

6

NUMERICAL SIMULATIONS

A numerical model was built with solid elements from the ABAQUS library of finite elements

(Figure 2a). The elements chosen for the columns were the C3D20RT while for the rest of the

surrounding structure were the C3D8RT. The C3D8RT is a 8-node while the C3D20RT is a 20

node linear finite element with reduced integration, an hourglass control solid element and a first-

order (linear) interpolation. These elements have one integration point, three degrees-of-freedom

per node corresponding to translations and six stress components in each element output.

The finite element mesh was generated automatically by the ABAQUS program and the side of the

finite elements was 50 mm in the specimen, walls and upper beams of the restraining frame and

100mm in the columns of the restraining frame (Figure 2b).

a) b)

Figure 2. Three-dimensional model and finite element mesh for the restraining frame and specimen

a) steel column embedded on a wall b) isolated bare steel column

The thermal and mechanical properties at high temperatures of the concrete were defined according

to EN 1992-1-2(2004) and of the steel according to EN1993-1-2(2005). For the bricks the

properties were not considered varying with temperature due to a lack of data available in the

literature of the specialty. Values used in the software Ozone, from the University of Liège, were

adopted (Cadorin, 2003).

On the un-exposed side, a convection coefficient of 4 Wm2/ºC and emissivity coefficients of 0.7 for

the concrete and 0.8 for the steel, and on the exposed side a convection coefficient of 25 Wm2/ºC

and an emissivity of 0.7 were used for both materials.

CASES STUDY

Table 1 and figure 3 describe the cases study in the numerical simulations and the correspondent

fire resistance test. The specimens were made of HEA160 and HEA200 profiles, 3m tall, with end

steel plates of 450mm x 450mm x 30mm, all of steel grade S355. The reference “Exx” indicates the

tests carried out on columns embedded on walls while the reference “Iyy” indicates the tests carried

out on isolated bare steel columns.

They were considered walls with the thickness nearly the same or smaller than the columns width or

height, depending on the orientation of the buckling axis in relation to the wall surface. The walls

thicknesses were chosen according to the commercial dimensions of the bricks.

Two different orientations of the profile in relation to the wall surface were considered: web parallel

and perpendicular to the walls surface. The reason for this choice is that a different behaviour was

expected, since in the two cases, bending occurred around the weak or strong axis of the cros

sections.

Test Steel profile Web in relation

wall surface

E03 HEA 200 parallel

E04 HEA 200 perpendicul







I16 HEA 160 No walls

I20 HEA 200 No walls

Figure 3 presents also the location of the thermocouples in the cross

numerical tests.

Figure 3. Cases study and location of thermocouples in the cross

Figure 4a) to d) presents a view cut

a) b) c) d)

Figure 4. Model of the FE mesh of the specimen

expected, since in the two cases, bending occurred around the weak or strong axis of the cros

Table 1. Cases Study

in relation to the

wall surface

Wall thickness

(mm)

Serviceability load

(kN)

parallel 180 1088

perpendicular 180 1088

parallel 140 704


parallel 140 1088


parallel 100 704


No walls - 704

No walls - 1088

Figure 3 presents also the location of the thermocouples in the cross-section in the experimental and

n of thermocouples in the cross-section of the columns.

of the cases study.

a) b) c) d)

mesh of the specimens a) Test E03 b) Test E04 a) Test E08

expected, since in the two cases, bending occurred around the weak or strong axis of the cross

Slenderness

42.2

42.2

52.8

52.8

42.2

42.2

52.8

52.8

52.8

42.2

section in the experimental and

section of the columns.

Test E08 a) Test E09

RESULTS

Thermal bowing

In Figure 5 a) and b) the behavior of the steel columns embedded on walls is observed. In the

beginning, the displacement is towards the side of the furnace, i.e., the side of the thermal action,

after which there is an inversion causing the column to move to the opposite side. Also, rotations on

top and bottom of the columns suffer the corresponding inversion. In figure 5c) a view of the

deformed shape of an isolated HEA200 is shown.

a) b) c)

Figure 5. Deformed shape of the columns a) Test E05 – instant t=440s (scale factor=10) b) Test

E05 – instant t=1445s (scale factor=5) c) Test I20 – instant=468s (scale factor=5)

Temperatures

In figure 6 the evolution of temperatures in several points (according to Figure 3), are depicted, at

mid-height of the steel columns. It is observed a very uniform temperature distribution in the

isolated column (test I16) and a great thermal gradient in the column embedded on walls (test E04).

(a) Test I16 (b) Test E04

Figure 6. Evolution of temperatures along the time at mid-height of the columns

Restraining forces

Figure 7 presents the variation of the restraining forces for steel columns HEA200 and HEA160

embedded on walls and isolated in function of the time. The values are refered to the initial applied

load (serviceability load). The fire resistance is here defined as the instant when the restraining

forces reach again the initial applied load.

0

100

200

300

400

500

600

700

0 2 4 6 8

Te

mp

era

ture

(ºC

)

t (min)

T2

T4

T6

T1

T5

T3

0

100

200

300

400

500

600

700

0 6 12 18 24 30

Te

mp

era

ture

(ºC

)

t (min)

T2

T4

T6

T1

T5

T3

(a) HEA 200 (b) HEA 160

Figure 7. Restraining forces in function of time

It can be clearlly observed in Figures 7 (a) and (b) that steel columns embedded on walls led to

higher fire resistances. Also, it is observed that the variation of restraining forces was lower than in

bare steel columns.

For the HEA200 columns, the embedding on walls, increase the fire resistance up to 21.4 minutes

(test E03) instead of 8.3 minutes for bare steel column (test I20). Also, it is observed that the

increasing of restraining forces can be as low as 0.4% for test E04, instead of 4% for the bare steel

column (test I20), i.e., 10 times lower (Figure 7 (a)).

For the HEA160 columns, a fire resistance of around 20 minutes was achieved for tests E05 and

E06 for columns embedded on walls. In therms of restraining forces, the lowest value observed is

1% for test E06 (Figure 7 (b)). This result was expected, once that both in tests E04 and E06, the

steel profile was completely embedded on the walls.

Tests E11, E06, E09 and E04 on columns embedded on walls and with the web perpendicular to the

wall surface provided lower restraining forces.

Tests E04 and E06, where columns were completely embedded on walls, presented higher fire

resistance, showing the beneficial effect of the walls on protecting the columns.

The behaviour in tests E05 and E10 are very similar to the test with the bare steel column (Figure 7

(b)). The failure seem to occur by buckling due to an abrupt reduction in the axial displacements

(indicating thermal bowing).

Tests E04 and E06 presented the most smooth evolution of restraining forces, once the columns are

almost totally embeded on walls, except the outter surface of the flange.

Axial Displacements and Lateral Deflections

Figures 8 to 11 present the evolution of the axial displacements and lateral deflections at midheigth

of the columns, in function of the time, for steel columns HEA200 and HEA160 embedded on walls

and isolated. In these graphs, positive axial displacements mean upwards and negative downwards.

Positive lateral displacements mean towards the side of the fire and negative to the oposite side.

The axial displacements measured on isolated bare steel columns were higher than the ones

measured on the steel columns embedded on walls. Thermal bowing is observed in both cases,

0,90

0,92

0,94

0,96

0,98

1,00

1,02

1,04

1,06

0 5 10 15 20 25 30

N/N

bR

d,2

0

t (min)

I20 E04 E03

E08 E09

0,90

0,92

0,94

0,96

0,98

1,00

1,02

1,04

1,06

0 5 10 15 20 25 30

N/N

bR

d,2

0

t (min)

I16 E05 E06

E10 E11

characterised by a slow deflection towards the fire and afterwards a suden inversion to the oposite

side. Failure occurs due to a sudden vertical displacement acompained by a big lateral deflection.

(a) E03 (wall thicness=180mm) (b) E08 (wall thicness=140mm)

Figure 8. Axial displacements and lateral deflections for isolated and embedded on walls HEA200

steel columns - web parallel to the wall surface

In Figure 8 it can be observed that the axial displacements were practically the same in both tests

(about 5mm). The column in test E08 after reach the value of the maximum axial displacement

presented a sudden decreasing confirming a deformation by buckling. The lateral deflections

towards the side of the fire where higher for specimens with a thinner wall.



steel columns - web perpendicullar to the wall surface

In Figure 9 both the axial displacements and the lateral deformations were very similar in tests E04

and E09. However the axial displacements in this case were nearly a half of the ones registered on

-10

-8

-6

-4

-2

0

2

4

6

8

10

0 5 10 15 20 25

Dis

pla

cem

en

t (m

m)

t (min)

Axial Displ - I20

Lateral Deflect - I20

Axial Displ - E03

Lateral Deflect - E03

-10

-8

-6

-4

-2

0

2

4

6

8

10

0 5 10 15 20

Dis

pla

cem

en

t (m

m)

t (min)

Axial Displ - I20


Axial Displ - E08


-10

-8

-6

-4

-2

0

2

4

6

8

10

0 5 10 15 20 25 30

Dis

pla

cem

en

t (m

m)

t (min)

Axial Displ - E04


Axial Displ - I20


-10

-8

-6

-4

-2

0

2

4

6

8

10

0 5 10 15 20

Dis

pla

cem

en

t (m

m)

t (min)

Axial Displ - I20


Axial Displ - E09


columns with the web parallel to the wall surface (Figure 8). This may be justified by the mass of

steel directly exposed to the heating source.

For the case of the test E08 and E09 the inversion of the lateral deflections and the axial

deformations occurred for different instants of time showing with this some influence of the wall

thickness on the behaviour of these columns.



steel columns - web parallel to the wall surface

In Figure 10 the lateral deflections and the vertical displacements were similar between them.

a) E06 (wall thicness=140mm) b) E11 (wall thicness=100mm)


steel columns - web perpendicullar to the wall surface

In Figure 11 the shape of the curves are very similar to the ones observed in the corresponding tests

for columns HEA200, E04 and E09. Once again the axial displacements in this case were nearly a

half of the ones registered on columns with the web parallel to the wall surface (Figure 10).

-10

-8

-6

-4

-2

0

2

4

6

8

10

0 5 10 15 20 25

Dis

pla

cem

en

t (m

m)

t (min)

Axial Displ - I16


Axial Displ - E05


-10

-8

-6

-4

-2

0

2

4

6

8

10

0 5 10 15 20

Dis

pla

cem

en

t (m

m)

t (min)

Axial Displ - I16


Axial Displ - E10


-10

-8

-6

-4

-2

0

2

4

6

8

10

0 5 10 15 20 25 30

Dis

pla

cem

en

t (m

m)

t (min)

Axial Displ - I16


Axial Displ - E06


-10

-8

-6

-4

-2

0

2

4

6

8

10

0 5 10 15 20 25 30

Dis

pla

cem

en

t (m

m)

t (min)

Axial Displ - I16


Axial Displ - E11


The inversion of lateral deflections and axial deformations, as registered in

for different instants of time showing with this some influence of the wall thickness on the

behaviour of the columns.

The specimens with columns with th

(th=140mm for HEA200 and th=100mm for HEA160), respectivelly tests

(b) and 11 (b), showed an abrupt decay of the vertical displacement

displacement reaches again the initial position. This fact may be

terms of fire resistance.

Fire Resistances

Table 3 presents the fire resistance

that the columns in contact with walls have a higher fire resistance than the ba

Furthermore, it is observed that the orientation of the web has no major influence in the

resistance. In fact, the fire resistance is

perpendicular to the wall surface but the differences are very small. The main influence is provide

by the thickness of the walls: thicker walls provided

Table 3. Fire resistance of the Test columns

HEA 160

Web

parallel

to the wall

Web

perpendicular

to the wall

Thick wall 18.8 20.2

Thin wall 16.3 9.4

COLUMNS AFTER

Figure 12 presents the deformed shape of

and (b) show HEA200 columns with the web perpendicular

with thick (th=180mm) and thin (th=100mm)

HEA160 columns with the same orientations of the web in relation to the walls, embedded

walls (th=140mm). In Figures 12 (e

HEA200 steel columns after fire.

(a)E04 HEA 200 (b)E08 HEA 200 (c

Figure 12. Columns after fire resistance tests

This study showed that steel columns

totally engulfed in fire. This means that the beneficial effect of the insulation provided by the walls

The inversion of lateral deflections and axial deformations, as registered in Figure 9, also occurred


with the web perpendicular to the wall surface and

(th=140mm for HEA200 and th=100mm for HEA160), respectivelly tests E09 and E11

an abrupt decay of the vertical displacement some mi

the initial position. This fact may be considered to be favourable in

presents the fire resistance obtained in this study for the columns. The main conclusion is

that the columns in contact with walls have a higher fire resistance than the bare

observed that the orientation of the web has no major influence in the

resistance is in general slightly higher for columns with the web

but the differences are very small. The main influence is provide

e walls: thicker walls provided greater fire resistances.

est columns (minutes)

HEA 200 HEA160

bare steel

HEA200

bare steelperpendicular

to the wall

Web

parallel

to the wall

Web

perpendicular

to the wall

21.4 15.4 7.7 8.

14.9 8.7

COLUMNS AFTER FIRE RESISTANCE TESTS

Figure 12 presents the deformed shape of some columns after the fire resistance tests.

with the web perpendicular and parallel to the wall surface

thin (th=100mm) wall respectivelly. Figures 12 (c

ame orientations of the web in relation to the walls, embedded

e) and (f) are presented the deformed shape of bare

c)E06 HEA 160 (d)E05 HEA160 (e) I16 HEA160

Columns after fire resistance tests

CONCLUSIONS

steel columns embedded on walls presented higher fire resistanc

fire. This means that the beneficial effect of the insulation provided by the walls

Figure 9, also occurred


e web perpendicular to the wall surface and thin wall

9 and E11, Figures 9

ome minutes after the

considered to be favourable in

The main conclusion is

re steel columns.

observed that the orientation of the web has no major influence in the fire

columns with the web

but the differences are very small. The main influence is provided

HEA200

bare steel

8.3

columns after the fire resistance tests. Figures 12 (a)

to the wall surface tested

c) and (d) show

ame orientations of the web in relation to the walls, embedded on thick

are presented the deformed shape of bare HEA160 and

(f) I20 HEA200

higher fire resistance than those

fire. This means that the beneficial effect of the insulation provided by the walls

plays a major influence over the detrimental effect of the thermal gradients developed in the column

cross-section.

In all situations under test, thermal bowing was observed, causing an inversion of lateral deflections

from the hot to the cold side of the wall. This behaviour led to a failure mode by bending, instead of

buckling. Buckling was only observed in bare steel columns.

The slenderness of the columns did not influence strongly the fire resistance, (for the tested

columns). Although, the higher was the slenderness the lower was the fire resistance.

The main parameter that influenced the behavior of the columns subjected to fire was the wall

thickness: thinner walls provided lower fire resistances.

The walls were also effective in preventing the columns with the web perpendicular to the wall

surface to fail around the minor axis. In these cases, the detrimental effect of the thermal bowing

seems to be canceled by the fact that failure is forced to occur around the strong axis. Moreover, no

local buckling was observed in the flanges due to the contact with the walls.

In experimental tests was observed some detachment of the columns in relation to the walls. This

reduces the fire resistance of the columns. The problem could be solved by welding some steel

connectors to the columns that will enhance the connection between the columns and the walls.

The fire resistance of the columns embedded on walls could also be enhanced by mounting bricks

between the flanges of the profiles.

Acknowledgement

The authors acknowledge the Portuguese Foundation for Science and Technology - FCT - MCTES

for their support.

REFERENCES

Cadorin, Jean-François (2003), Compartment Fire Models For Structural Engineering, PhD Thesis,

Faculté de Sciences Appliqués, Université de Liège.

Correia, A.M., Rodrigues J. P., Silva, V. P., (2007), Studies on the fire behavior of steel columns

embedded on walls, Proceedings of 11th

International Conference on Fire Science and Engineering

- Interflam. London. UK., 641-652.

Correia, A.M., Rodrigues J. P., Silva, V. P. (2009), Experimental Research on the Fire Behavior of

Steel Columns Embedded on Walls, Proceedings. of International Conference Applications of

Structural Fire Engineering, Prague, Czech Republic, 417-422.

Correia, A.M., Rodrigues J. P., Silva, V. P., Laim, L. (2009), Section Factor and Steel Columns

Embedded in Walls, Proceedings. of 11th

Nordic Steel Construction Conference, Malmö, Sweden.

EN1993-1-1(2005). “Design of steel structures – part 1-1: General Rules and rules for buildings”,

CEN, Brussels.

EN1993-1-2(2005). “Design of steel structures – part 1-2: General Rules – Structural fire design”, CEN, Brussels.

EN1992-1-2(1995). “Design of concrete structures – part 1-2: General Rules – Structural fire

design”, CEN, Brussels.

EN1992-1-2(1995). “Design of concrete structures – part 1-2: General Rules – Structural fire

design”, CEN, Brussels.

EN1991-1-2(2002). “Eurocode 1: Actions on structures – part 1-2: General actions – Actions on

structures exposed to fire”, CEN, Brussels.

Documents

NUMERICAL STUDY ON THE BEHAVIOR OF STEEL COLUMNS EMBEDDED ON BRICK WALLS SUBJECTED TO FIRE