A Multi-Objective Advanced Design Methodology of … · A Multi-Objective Advanced Design Methodology of Composite Beam-to-Column Joints Subjected to Seismic and Fire Loads Raffaele

A Multi-Objective Advanced Design Methodology of Composite Beam-to-Column Joints Subjected to Seismic and Fire Loads

Raffaele Pucinottia, Fabio Ferrariob, Oreste S. Bursib

a Department of Mechanics and Materials, Mediterranean University of Reggio Calabria, loc. Feo di Vito, Reggio Calabria, 89126, Italy

b Department of Mechanical and Structural Engineering, University of Trento, via Mesiano 7, Trento, 38050, Italy

Abstract. A multi-objective advanced design methodology dealing with seismic actions followed by fire on steel-concrete composite full strength joints with concrete filled tubes is proposed in this paper. The specimens were designed in detail in order to exhibit a suitable fire behaviour after a severe earthquake. The major aspects of the cyclic behaviour of composite joints are presented and commented upon. The data obtained from monotonic and cyclic experimental tests have been used to calibrate a model of the joint in order to perform seismic simulations on several moment resisting frames. A hysteretic law was used to take into account the seismic degradation of the joints. Finally, fire tests were conducted with the objective to evaluate fire resistance of the connection already damaged by an earthquake. The experimental activity together with FE simulation demonstrated the adequacy of the advanced design methodology.

Keywords: Earthquake, Steel-concrete composite joint, full strength joints, seismic actions, fire actions, FE model.

INTRODUCTION

Major earthquake in urban areas have often been followed by significant conflagration that have been difficult to control and have resulted in extensive damage to property. Major contributing factors have been identified as accidental ignition due to earthquake shaking, external fire spread through vegetation and inadequate building separation, earthquake damage to building’s fire safety systems, loss of water supplies for fire suppression, and the lack of intervention by fire fighters due to inadequate resources and obstructed access to the site of the fire. Earthquake, then, increases the risk of loss of life if a fire occurs within the building. It is obvious therefore that fire after earthquake is a scenario that should be properly addressed in any performance base design in locations where significant earthquake can occur. The science around predicting the likelihood of such events is so complex that is not considered a design requirement; however given that an important part of Europe is a seismic prone area, the combination of a fire after an earthquake poses a very real danger. Fire and earthquake are accidental actions and have been treated most often as independent events, which each considered separately. However, if the occurrence of a fire does

not increase the probability of an earthquake, the opposite is not true. As Kobe earthquake tragically demonstrated in 1995, a city can be devastated by numerous fires following the earthquake. Post earthquake fire is a low probability event but with high potential consequences, and it is appropriate that it be included in the risk analysis. According to the modern seismic codes, ordinary structures are designed to suffer damage to some extent during strong earthquake, exploiting the structure own ductility to avoid collapse and safeguard human lives. Then, a fire coming soon after an earthquake will find a different, more vulnerable, structure with respect to the initial, undamaged, one. Depending on the extent of damage, the fire resistance rating of the structure could be significantly reduced.

In this paper a multi-objective advanced design methodology dealing with seismic actions followed by fire on steel-concrete composite full strength joints with concrete filled tubes is proposed. The data obtained from experimental tests have been used to calibrate a model of joint and a hysteretic law was used to take into account the seismic degradation. Subsequently seismic simulations on several moment resisting frames were performed and fire tests were conducted on both undamaged and damaged specimens.

BEAM-TO-COLUMN JOINT

In order to obtain the values of the actions used in the design of the proposed joints, two moment resisting frames having the same structural typology but different slab systems were analyzed: a composite steel-concrete slab with structural profiled steel sheeting and a concrete slab composed of electro-welded lattice girders (Figure 1).

a) b) c) FIGURE 1. Relevant Plains, Structure typology and Elevation of five-storey MR Frame; a) slab with

profiled ribbed steel sheeting; b) slab with prefabricate lattice girder, c) elevation.

The composite steel-concrete office-building had 5 floors with 3.5 m storey height. It was made up by three moment resisting frames placed at the distance of 7.5 m each in the longitudinal direction, while it was braced in the transverse direction. In particular, the main moment resisting frame is made up by two bays spanning 7.5 m and 10.0 m in the solution with steel sheeting with a distance between secondary beams equal to 2.5 m; and by two bays spanning 7.0 m and 10.5 m in the solution with lattice steel girders with a distance between secondary beams equal to 3.5 m. The main beams were IPE 400 while the secondary ones were IPE 300. The columns were concrete-filled column with a CFT filled tubular column steel profile with a diameter of 457 mm and a thickness of 12 mm. Buildings were designed to have a dissipative

structural behaviour according to point 7.1.2 of Eurocode 8 [4]. The effective width of the slab in composite beams was assigned according to Eurocode 4 (point, 5.4.1.2) [2, 3] for static and fire analyses and Eurocode 8 (point, 7.6.3 [4]) for seismic analyses. Beam-to-column connections were assumed to be rigid (5.1.2-EC4 [2]). Moreover, the influence of second-order effects on the determination of the action effects have been neglected since the criterion given in 5.2.1(4)-EC3 [1]. This allowed also to neglect member imperfections in composite compression members within global analysis. The welded joints were made by two horizontal diaphragm plates split into two equal halves along the diagonal for fabrication convenience and easiness of assembling. Once each side were properly placed on the pipe the two halves are attached with a full-joint penetration groove weld. In detail, the inferior plates are welded to column in the shop, while the superior plates are welded on site as should be understood from Figure 2.

230

60

475 300 475

625 6251250

40R22

8,5

290

= 290 x 1250 x 16A 5A

Horizontal plates

Vertical plate

300

700

= 300 x 700 x 14A 6

1250

700

A 6= 300 x 700 x 14

121,5121,530

0

369

1616

300

1950

Joint

R228,5

4,3

14 18

23°

23° 23° 23° 23°33

°

1814

4,3R22

8,5

Stiffeners

20x2

020

0

150

50

100 25

125= 125 x 200 x 12

A 4

plate

FIGURE 2. Detail of welded steel-concrete composite joint. (Dimensions in mm).

EXPERIMENTAL PROGRAMME

The experimental programme concerned the execution of 6 tests on full-scale substructures representing the interior welded beam-to-column joint in seismic conditions (Table 1). Specimens (Figure 3) were designed and fabricated according to Eurocode 4 [2, 3] and Eurocode 8 [4] provisions. In all composite specimens the connections between steel beam and desk were made by Nelson 19 mm stud connectors with an ultimate tensile strength fu=450 MPa. Additional Nelson 19 mm studs were localized around the column in order to enforce a better connection between the column and the composite slabs (Figure 2). Experimental tests were carried out at the Laboratory for Materials and Structures of the University of Trento. The scheme of the experimental set-up was detailed in [6] together with the employed instrumentation, while the geometric property specimens are shown in the Figure 3. The joint specimens were subjected to monotonic and cyclic loadings up to collapse, according to the ECCS stepwise increasing amplitude loading protocol, modified with the SAC procedure [4]: ey=17.5 mm. All specimens exhibit a good performance in terms of resistance, stiffness, energy dissipation and ductility. Both the overall force-displacement relationships and the moment-rotation relationships relevant to plastic hinges formed in the composite beams exhibit a hysteretic behaviour with large energy dissipation without evident loss of resistance and stiffness (Figures 4, and 5). From the

results of monotonic tests, it was observed that the joints exhibited elasto-plastic behaviour with limited degradation. There was practically no difference in the monotonic responses between the joint with steel sheeting slab and joint with prefabricate slab.

A

150

A

IPE 400

CFT

Ø 4

57 x

12

a) CFT

Ø 4

57 x

12

A

A

IPE 400

b)

FIGURE 3. Welded Beam-to-Column joint specimens; a) endowed with steel sheeting; b) endowed with prefabricate girder

Specimen WJ-P1

-1000

-600

-200

200

600

1000

-14 -10 -6 -2 2 6 10 14

Displacement [e/ey]

Forc

e [k

N]

Specimen WJ-P1

-800-600-400-200

0200400600800

1000

-80 -60 -40 -20 0 20 40 60 80

φ [rad]

M [k

Nm

m]

Left HingeRight Hinge

Specimen WJ-P2

-1000

-600

-200

200

600

1000

-14 -10 -6 -2 2 6 10 14

Displacement [e/ey]

Forc

e [k

N]

Specimen WJ-P2

-800-600-400-200

0200400600800

1000

-80 -60 -40 -20 0 20 40 60 80

φ [rad]

M [k

Nm

m]

Left HingeRight Hinge

Specimen WJ-S1

-1000

-600

-200

200

600

1000

-14 -10 -6 -2 2 6 10 14

Displacement [e/ey]

Forc

e [k

N]

Specimen WJ-S1

-800-600-400-200

0200400600800

1000

-80 -60 -40 -20 0 20 40 60 80

φ [rad]

M [k

Nm

])

Left Hinge

Right Hinge

Specimen WJ-S2

-1000

-600

-200

200

600

1000

-14 -10 -6 -2 2 6 10 14

Displacement [e/ey]

Forc

e [k

N]

Specimen WJ-S2

-800-600-400-200

0200400600800

1000

-80 -60 -40 -20 0 20 40 60 80

φ [rad]

M [k

Nm

]

Left HingeRight Hinge

FIGURE 4. Cyclic Test results of Specimen WJ: Force vs. Displacement and Moment vs. Rotation

curves

TABLE 1. Experimental programme. No Name Test

Method Type of Specimen

1 WJ-P1 Cyclic Specimens with electro-welded lattice girders slab and no Nelson connectors around the column, without reinforcement plates

2 WJ-P2 Cyclic Specimens with electro-welded lattice girders slab and Nelson connectors around the column, without reinforcement plates

3 WJ-S1 Cyclic Specimens with electro-welded lattice girders slab and no Nelson connectors around the column, without reinforcement plates

4 WJ-S2 Cyclic Specimens with profiled Steel Sheeting slab and Nelson connectors around the column, without reinforcement plates

5 WJ-PM Monotonic Specimens with profiled Steel Sheeting slab and no Nelson connectors around the column, without reinforcement plates

6 WJ-SM Monotonic Specimens with profiled Steel Sheeting slab and no Nelson connectors around the column, without reinforcement plates

Specimen WJ-PM

0

200

400

600

800

1000

0 6 12 18 24 30Displacement [e/ey]

Forc

e [k

N]

Specimen WJ-PM

-1000

-600

-200

200

600

1000

-160 -120 -80 -40 0 40 80 120 160

φ [mrad]

M [k

Nm

]Right Plastic Hinge

Lift Plastic Hinge

Specimen WJ-SM

0

200

400

600

800

1000

0 6 12 18 24 30Displacement [e/ey]

Forc

e[kN

]

Specimen WJ-SM

-1000

-600

-200

200

600

1000

-160 -120 -80 -40 0 40 80 120 160

φ [mrad]

M [k

Nm

]

Right Plastic HingeLift Plastic Hinge

(-M)

FIGURE 5. Specimen WJ: Force vs. Displacement and Moment vs. Rotation curves

COMPUTER FE MODELS

FE Model of the Specimens

The data obtained from monotonic and cyclic experimental tests have been used to calibrate a model of the joint in order to perform seismic simulations on several moment resisting frames. The model used to perform these analyses is based on two parallel springs at the end of the beam connected to a rigid panel which represents the rigid connection with the column. Each spring is used to match the properties of the beam under sagging and hogging bending moment. Another spring is used to connect this panel with the shear panel of the column (see Figure 6). The main purpose of this spring is to take into account the shear deformation of the connection in order to improve the calibration of the model.

The calibration of the springs used in the model, has been done twice for each type of slab, one for the experiments performed without connectors in the upside of the top plate of the connection and other for the experiments performed with connectors. IDARC program [9] was used to perform the simulation. A hysteretic law has been used to take into account the seismic degradation of the joint according to a modification of the Bouc-Wen model implemented by Silvaselvan-Reinhorn in IDARC [10]. The real measured properties of the concrete and steel have been used in the computer model in order to match as accurately as possible the results of the experimental tests [6]. The behaviour has been validated in two ways: a) the amount of energy dissipated by the whole model and by the plastic hinge of the composite beam during the computer simulations have to match the energy dissipated during the laboratory tests by the same elements; b) the hysteretic behaviour exhibited in the computer simulations has to be similar to that shown in the laboratory tests.

a)

473

236,

5

198,25

99,125 99,125

236,

5

457

99,125 99,125

b) FIGURE 6. a) Computer model of specimens; b) Joint details of computer model

The hysteretic behaviour exhibited in the computer simulations has to be similar to that shown in the laboratory tests (Figure 7). Once the behaviour of the joints have been calibrated, two frames have been modelled in IDARC [9] for both type of slabs in order to evaluate its dynamic properties. In the application the Park and Ang [7] damage index modified by Chai and Romstad [8] was considered.

Confronto Forze Spostamenti Martinetto

-800

-600

-400

-200

0

200

400

600

800

-2.50E+02 -2.00E+02 -1.50E+02 -1.00E+02 -5.00E+01 0.00E+00 5.00E+01 1.00E+02 1.50E+02 2.00E+02

delta (mm)

F (k

N)

IDARC Sperimentale

Confronto Cerniera Plastica Sinistra

-1000000

-800000

-600000

-400000

-200000

0

200000

400000

600000

800000

1000000

1200000

-6.00E-02 -4.00E-02 -2.00E-02 0.00E+00 2.00E-02 4.00E-02 6.00E-02 8.00E-02

Ø (rad)

M (k

Nm

m)

FIGURE 7. a) V- δ curve of the whole mobeam, interior joint,

FE Mode

The structural behaviour under vbeen simulated numerically, utilizing(Figure 1). Incremental dynamic ana(two frames for each type of slab). Thperformed with each frame. The acc

IDARC Exp

IDARC Sperimentale del, interior joint, prefabricated slab; b) M-ø curve of the prefabricated slab. Sx plastic hinge

l of the Structures

ertical static loads and under seismic loads has a bi-dimensional model for the principal frame lyses have been carried out with the four frames ree incremental time history analyses have been

elerograms used to perform the simulation have

been artificially generated to match Type 1 response spectrum and soils type A, type B and type D ( Figure 8a). Each accelerogram has been used to perform twenty time history analyses with different peak ground accelerations ranging from 0.1g to 2.0g with an interval of 0.1g, thus resulting in an incremental dynamic analysis. Additional simulations were also done to identify the peak ground acceleration when the first yielding occurred and when the 35 mrad plastic rotation was reached.

0.00

0.20

0.40

0.60

0.80

1.00

1.20

1.40

1.60

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0T [s]

Sa/g

TYPE DTYPE BTYPE A

-0.50

-0.40

-0.30

-0.20

-0.10

0.00

0.10

0.20

0.30

0.40

0.50

0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0

T [s]

a/g

FIGURE 8. a) Response spectrums of accelerograms matching soil type A, B and D of Type 1 EC8 spectrum.; b) Time history of artificial accelerogram matching soil type A of Type 1 EC8 spectrum.

The behaviour factor has been calculated using the equation:

y

udyn pga

pgaq = (1)

where pgau corresponds to the peak ground acceleration that induces a first plastic rotation in a beam-to-column joint of about 35 mrad; while pgay corresponds to the peak acceleration that produces the first yielding in the structure.

Capacity curves and the main behaviour factors obtained from IDA analysis are shown in Figure 9 and in the Table 2 respectively.

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

0.00% 0.50% 1.00% 1.50% 2.00% 2.50% 3.00% 3.50% 4.00%Drift [%]

V/W

1_storey2_storey3_storey4_storey5_storeyFrame

FIGURE 9. IDA capacity curves, with soil Type D, for steel sheeting frame without connectors

TABLE 2. Behaviour factors from IDA analysis Without Connectors With Connectors IDA A IDA B IDA D IDA A IDA B IDA D

Steel sheeting slab 3.82 3.57 3.58 4.24 4.10 4.00 Prefabricated slab 3.17 3.42 3.48 3.75 3.94 5.03 Largest values of the damage index for the all studied frames, under a seismic

excitation of 0.4 g, are 0.503 for external joints and 0.419 for internal joints. The damage in joints is limited and repairable [12].

FIRE RESPONSE

In order to accurately simulate the damage, calculated in the previous session, due to a seismic event, a specified deformation was applied on the specimens. In particular, only one of each type of connections (Steel sheeting slab and prefabricated lattice slab) was subject to the damage test while all specimens were fire tested.

The purpose of the fire tests was to determine the residual fire resistance of a damaged connection subject to fire limit state loading. Fire tests were conducted, at Cardington by BRE, with asymmetric loading on the internal connections representing the load in the column of the frame from primary beams of different length (Figure 1).

Fire test results

Undamaged Specimen endowed with steel sheeting slab. Figure 10a shows both the west and the east load applied and the average atmosphere temperature evolution, while Figure 10b shows the atmosphere temperatures relative to the standard fire curve; figure 10c shows instead the temperature distribution on the composite beam.

The fire test led to full depth cracking of the slab, runaway deflection (Figure 10 d). The test was terminated due to runaway deflection after approximately 40 minutes.

Steel Sheeting Slab-Undamaged

0

50

100

150

200

0 20 40 60 80 100Time [min]

Loa

d (K

N)

Tem

p. [°

C]

West LoadEastLoadTemp.

500

1000

1500

0

2000

a)

Steel Sheeting Slab-Undamaged

0

200

400

600

800

1000

1200

0 20 40 60 80 100Time [Min]

Tem

p. [°

C]

G3G4G1G2SFCAv. Atmos

Burner

b) Steel Sheeting Slab-Undamaged

0100200300400500600700800

0 20 40 60 80 100Time [Min]

Tem

p. [°

C]

TC7TC8TC9TC10TC11TC12TC13Av. Atmos

Steel Sheeting Slab-Undamaged

-40

-20

0

20

40

60

80

0 200 400 600 800

Temp. [°C]

Ver

tical

Def

lect

ion

[mm

]

C1061C1062C1063C1064

+VE Denotes downwards deflection

c) d) FIGURE 10. Specimen endowed with Steel sheeting slab, a) The west and the east load applied and the

average atmosphere temperature b) Atmosphere Temperatures; c) East Beam Temperature; d) Temperature vs. Vertical Deflection

Damaged Specimen endowed with steel sheeting slab. The test had to be terminated after approximately 34 minutes due to runaway deflection. The atmosphere temperature is shown in Figure 11a while Figure 11b shows the measured column temperatures. Following the fire test the profiled steel sheeting had separated from the slab, the slab had cracking both along the surface and through the depth of the slab and there was extensive buckling of both the lower flange and the web of the composite beam in the area of the connection.

Undamaged Specimen endowed with prefabricated lattice slab. For this specimen the test continued through to 60 minutes. At this points the rate of deflection was

increasing but had not yet reached the point where it was not possible to maintain the applied load. There was no permanent deformation and no sign of any significant damage from the fire test. The critical temperatures and the temperature-deformation chracteristics are shown in Figure 12.

Steel Sheeting Slab-Damaged

0

200

400

600

800

1000

1200

0 10 20 30 40 50Time [Min]

Tem

p. [°

C]

Serie1

Serie2

Serie3

Serie4

SFC

Burner

a)

Steel Sheeting Slab-Damaged

0100200300400500600700800

0 5 10 15 20 25 30 35 40 45 50 55Time [Min]

Tem

p [°

C]

TC1TC2TC3TC4TC5TC6

b) FIGURE 11. Damaged Specimen endowed with Steel sheeting slab, a) Atmosphere Temperatures; b)

Column Temperature;

Prefabricated Slab-Undamaged

0

50

100

150

200

0 20 40 60 80 100

Time [min]

Loa

d (K

N)

Tem

p. [°

C]

West LoadEastLoadTemp.

500

1000

1500

0

2000

a)

Prefabricated Slab-Undamaged

0

200

400

600

800

1000

0 20 40 60 80 100 120Time [Min]

Tem

p. [°

C]

C1021C1022C1023C1024C1025C1026C1027

b) FIGURE 12. Undamaged Specimen endowed with Prefabricated slab, a) The west and the east load

applied and the average atmosphere temperature b) East Beam Temperature

Damaged Specimen endowed with prefabricated lattice slab. The behaviour of the prefabricated lattice slab specimen which had been subject to the damage test was very similar to that for the undamaged specimen.

Steel Sheeting Slab

0

200

400

600

800

1000

0.00 2.00 4.00 6.00 8.00 10.00Φ [mrad]

Hog

ging

Mom

ent

[kN

m]

temp=20°CAbaqus - after 30' of exposure Experimental - after 30' of exposure - UndamagedExperimental - after 30' of exposure - Damaged

a) b) FIGURE 13. Specimen endowed with Steel Sheeting Slab a) Comparison among the resistance of

undamaged and damaged joint estimated at 20°C and after 30’of fire exposition ; b) Abaqus FE model

With reference to the undamaged and damaged specimens endowed with steel sheeting slab, in Figure 13a the comparison among the resistance estimated at 20°C and after 30 minute is shown. In particular the value of the bending moments were calculated using the component method, considering both the experimental and the preliminary numerical temperature distribution in the joins. In this last case the Abaqus 6.4.1 code [11] was employed to conduct the thermal analysis [6]. The hogging capacity moment becomes approximately 38% of the initial value after 30 minute of fire exposition in the case of the undamaged specimen and 31% in the case of the damaged specimen. Instead, the capacity hogging moment, considering the

temperature distribution obtained by ABAQUS code (Figure 13b), becomes approximately 24% of the initial value.

CONCLUSIONS

A multi-objective advanced design methodology dealing with seismic actions followed by fire on steel-concrete composite full strength joints with concrete filled tubes is proposed in this paper. Due to the fact the joint are full-strength and rigid, the energy dissipating mechanisms in moment resisting frames with this type of joint will entirely rely on the formation of plastic hinges at the end of the beams. The study showed an optimum behaviour of the joint components till a damage index of Di=0.5. A behaviour factor of about 4 has been observed for the analyzed composite frames. The fire tests showed that there was no noticeable difference in the performance between the damaged and undamaged specimens; moreover these demonstrated the effectiveness of the strategy based on the protection of steel connection details by means of the exploitation of concrete both in the column and in the slab.

ACKNOWLEDGMENTS

This research is developed within the financial support of UE under the project PRECIOUS (RFS-CR-03034) which the writers are grateful.

REFERENCES

1. UNI EN 1993-1-1. 2005. Eurocode 3: Design of steel structures. Part 1: General rules and rules for buildings;

2. UNI EN 1994-1-1. 2005. Eurocode 4: Design of composite steel and concrete structures - Part 1.1: General rules and rules for buildings;

3. UNI EN 1994-1-2. 2005. Eurocode 4: Design of composite steel and concrete structures - Part 1.2: General rules – Structural fire design;

4. UNI EN 1998-1. 2005. Eurocode 8: Design of structures for earthquake resistance - Part 1: General rules, seismic actions and rules for buildings;

5. ECCS 1986. Recommended Testing Procedure for Assessing the Behaviour of Structural Steel Elements under Cyclic Loads. ECCS Publication n° 45;

6. Bursi O. S., Ferrario F., Pucinotti R., (2008) Steel-Concrete Composite full strength Joints with Concrete Filled Tubes: Design and Test Results, 12th International Symposium on Tubular Structures October 8 to 10, 2008 Shanghai, China;

7. Park Y. J. and Ang, A. H. S., (1985) Mechanistic Seismic Damage Model for Reinforced Concrete, Journal of Structural Engineering, vol.111 n.4;

8. Chai Y. H. and Romstad K. M., (1997) Correlation Between Strain-Based Low-Cycle Fatigue and Energy-Based Linear Damage Models, Earthquake Spectra, Vol. 13, No. 2, pp. 191-209;

9. Valles R.E., Reinhorn A.M., Kunnath S.K., Madan A., (1996) IDARC2D Version 4.0: A Program for the Inelastic Damage Analysis of Buildings. Technical Report NCEER-96-0010;

10. Silvaselvan M.V., Reinhorn A.M., (1999) Hysteretic Model for Cyclic Behaviour of Deteriorating Inelastic Structures, Technical Report MCEER-99-0018;

11. Hibbitt, Karlsson and Sorensen 2000. “ABAQUS User’s manuals”. 1080 Main Street, Paw-tucket, R.I. 02860.

12. Williams M. S., Sexsmith R. G., (1995), Seismic Damage Indices for Concrete Structures: A State-of-the-Art Review., Earthquake Spectra, 11, 2, 319-349]