ANALYSIS OF STEEL-CONCRETE COMPOSITE BEAM-TO-COLUMN JOINTS: WELDED SOLUTIONS Oreste S. Bursi Department of Mechanical and Structural Engineering, University of Trento Trento, Italy [email protected] Fabio Ferrario Department of Mechanical and Structural Engineering, University of Trento Trento, Italy [email protected] Raffaele Pucinotti Department of Mechanics and Materials, Mediterranean University of Reggio Calabria Reggio Calabria, Italy [email protected]

ABSTRACT A multi-objective design methodology dealing with seismic-induced fire on steel-concrete composite moment resisting frames endowed with concrete filled tubes full strength joints is presented in this paper. In order to achieve goals analytical and FE simulations including thermal analyses were carried out to design the proposed joints followed by experimental tests under monotonic and cyclic loadings. A total of six specimens was designed and subjected to lateral loads. The specimens were subassemblages of interior beam-to-column joints, connected by means of welded connections. Relevant experimental results are presented and commented. Furthermore, since the scope of the project was to promote joint typologies able to survive a seismic-induced fire, before being subjected to fire loadings specimens have been damaged by imposing monotonic loads equivalent to damage induced by seismic excitations. Based on experimental results valuable information were obtained about the performance of the proposed joint typology. Moreover, seismic and fire analyses have provided further information on the performance of moment resisting frames endowed with the joint typology. INTRODUCTION Major earthquakes in urban areas were often followed by significant conflagrations that were difficult to control and resulted in extensive damage to property. For instance in the 1995 Kobe earthquake, more than a few fires occurred in fire-resistive buildings (Kobe City Fire Dept., 1995). Also, various surveys revealed that many fire protection systems, e.g. sprinkler systems, were damaged by earthquakes and lost their proper function because of mechanical failure and/or deformation by earthquake motion (MFIAJ, 1995). As a result, on one hand seismicinduced fire risk assessment methods to evaluate fire risk according to size and type of

buildings, installed fire protection systems and intensity of input earthquake motion have been developed (Yashiro et al., 2000); on the other hand, some studies on the evaluation of the fire resistance rating reduction of perimeter moment-resisting multi-storey frames owing to seismic loading have been performed (Della Corte et al., 2003). At the beam-to-column joint level, different studies were carried out to evaluate the performance of steel joints under fire both from a numerical or a design viewpoint (Simoes da Silva et al., 2007, Simoes da Silva et al., 2005, Block et al., 2007). However, the research seems to be very limited with respect to the performance of beam-to-column joints under fire loading, even though research is still active in the field of monotonic and cyclic loading (Gil and Bayo, 2007 and 2008, Salvatore et al., 2005). This context of limited information on beam-to-column steel-concrete composite joints under fire and seismic-induced fire, motivated the European Union-financed research project PRECIOUS, devoted to the development of fundamental data, design guidelines and prequalification of two types of fire-resistant composite beam-to-column joints: one endowed with partially reinforcedconcrete-encased columns with I-section; one with concrete filled tubular columns with circular hollow steel section (Bursi et al., 2008). Because fire and earthquake are accidental actions and have been treated most often as independent events with each considered separately (EN Eurocode 1991-1-2, 2004, EN Eurocode 1998-1, 2005) the project fitted in a modern multiobjective performance-based design where fire safety is considered on a structure characterised by stiffness deterioration and strength degradation owing to seismic actions. In detail: the intensity measures were defined by means of artificial accelerograms and natural fires; engineering indices were defined in terms of joint distortions, interstorey drifts and time for fire resistance; damage measures owing to earthquake and fire were considered both at the joint and at the frame level taking into account the interaction between joints and members; decision variables were related to life safety (SEAOC, Vision 2000). The project achieved its goals through a balanced combination of analytical/numerical and experimental work. In this paper, we only concentrate on Type 2 joint, for which we present: i) the definition and the joint design for seismic and fire loadings; ii) the analysis of the mechanical and thermal behavior of this typology by means of experimental tests; iii) results from moment resisting frames endowed with the investigated joint solution under seismic, fire and induced-seismic fire. DESIGN OF REFERENCE FRAMES UNDER SEISMIC AND FIRE In order to obtain design actions on joints, two moment resisting frames have been designed with the same structural typology but different slab systems. In the frame design, particular attention was paid to the layout of structural elements according to economic criteria and to their optimization both for vertical and lateral loads. The reference structures are depicted in Figure 1(a)-1(c) and are made up of three moment resisting frames placed at the distance of 7.5 m each in the longitudinal direction; while they are braced in the transverse direction with an interstorey height of 3.5 m. Two different slab systems were considered: i) a concrete slab composed of electro-welded lattice girders as shown in Figure 2(a); ii) a composite steelconcrete slab with structural profiled steel sheeting as illustrated in Figure 2(b). All slabs were arranged in parallel to main frames. All connections between steel beams and slabs were full and made by Nelson 19 mm stud connectors with an ultimate tensile strength fu=450 MPa. In both cases, composite beams were realized with S355 IPE400 steel profiles, while composite columns were realized with 457 mm circular steel tubes with 12 mm thickness. The seismic performance of typical frames was evaluated by means of nonlinear static pushover analysis. Afterwards, the fire design issue was considered and the structural fire performance of the complete frame was evaluated by means of the SAFIR software (Franssen, 2000).

a)

b)

c)

Figure 1 – Geometric layout of reference structures: a) structure with slabs with prefabricated lattice girders; b) structure with slabs with profiled steel sheeting; c) frame elevation pos.C 8 Ø 16 / 200

pos.C 7 Ø 16 / 250

1500

pos.B 3+3 Ø 12

3Ø12

mesh HD 620

mesh HD 620 3Ø12 pos.B

560 720 880

2100 2000 1900

a)

300150

1Ø10 pos.D

160

pos.B 3+3 Ø 12

5080 5180

pos.A 4 Ø 12 / 100 pos.A 4 Ø 12 / 100 pos.C 7 Ø 16 / 250 mesh Ø 6 / 200x200

2100 2000 1900

1400

1Ø10 pos.H

150

1Ø10 pos.H 1Ø10 pos.D

400

1Ø8 pos.E HD 10/14/8 h=9,5cm

160

560 720 880

pos.A 5 Ø 12 / 100 mesh Ø 6 / 20x200

3Ø12

pos.A 5 Ø 12 / 100 pos.C 8 Ø 16 / 200

5080 5180

b)

Figure 2 – Schematic of typical cross sections: a) solution with prefabricated lattice girders; b) solution with steel sheeting. (Dimensions in mm) The slabs composed by prefabricated lattice girders illustrated in Figure 2(a) were endowed with 3+3φ12 longitudinal rebars and 5+5φ12@100 mm plus 8+8φ16@200 mm transversal steel bars. A mesh φ6@200x200 mm completed the slab reinforcement. Conversely, the slab with profiled steel sheeting was endowed with 3+3φ12 rebars, 4+4φ12@100 mm and 7+7φ16@250 mm transversal steel bars as shown in Figure 2(b); a mesh φ6@200x200 mm was adopted. We recall that longitudinal and transversal rebars are needed to activate strut and tie Mechanism 1 and 2 foreseen in EN Eurocode 1998-1 (2005); one couple of longitudinal rebars was designed to face damage owing to seismic loading. The rebar steel grade was S450B while the concrete class was C30/37. Composite beams were of Class 2. The concrete filled tubular (CFT) columns were S355 endowed with 8φ16 longitudinal rebars and φ8@150 mm stirrups as illustrated in Figure 3. The steel grade was S450B with concrete class was C30/37. pos.E Ø 8 / 15 cm

325

pos.D 8 Ø 16

120 150

2490

90

150 150

95

135

40

50

2670

Figure 3 - Column stub and reinforcements capable of hosting a through-column web plate Due to circular CFTs, the seismic design of composite beam-to-column joints was conceived to provide both adequate overstrength and stiffness with respect to connected beams, thus forcing

plastic hinges formation in adjacent beams (EN Eurocode 1998-1, 2005). As a result jointswere detailed by using the component method as shown in Figure 4a (EN Eurocode1993-1-8. 2005). Hogging Moment Resistance of Joint 1000

M [kNm]

800 600 400 200 0 0

0.5

1

1.5

2

2.5

3

3.5

4

[mrad]

a)

T=20°C

30' exposure

60' exposure

b)

Figure 4 – (a) Interior Type 2 joint and mechanical model; (b) Moment-rotation relationship of the joint as a function of the time fire resistance The following components were considered in the method: concrete slab in compression; upper horizontal plate in compression; vertical plate in bending and lower horizontal plate in tension, for sagging moment; reinforcing bars in tension, upper horizontal plate in tension; vertical plate in bending and lower horizontal plate in compression for hogging moment. The components concrete slab in compression and upper horizontal plate in tension were identified by means of FE models set with ABAQUS (Hibbitt et al., 2000) as illustrated in Figure 5 and 6, respectively.

a)

b)

Figure 5 – Assumed idealized mechanisms: a) Mechanism 1; b) Mechanism 2

a)

b)

Figure 6. FE model of a plate in tension: (a) elastic stresses; (b) inelastic stresses and width. The effective width b of plates was approximately assumed to be half of the total width plate. In the application of the component method the composite column was assumed to be infinitely

rigid. Beam-to-column joints were rigid full-strength joints with respect to adjacent beams satisfying the condition that M j , pl , Rd was at least 1.3 M b , pl , Rd (EN Eurocode 1998-1, 2005). Moreover, full groove and fillet welds were used. In detail, fillet welds satisfied the relationship:

Rd ≥ 1.1 ⋅ γ ov ⋅ R fy

(1)

with an overstrength factor γ ov = 1.25 . The beam-to-column joint is shown in Figure 7. It is composed of two horizontal plates and a vertical through-column plate. Only half of the top plate has to be welded on site. The presence of the vertical through column plate shown in the top left of Figure 7 and the presence of concrete close to the top plate help the joint to keep a lower temperature of these components to fire exposure. Thus the 15 min induced-seismic fire resistance objective should be achieved. Joint

Horizontal plates

475

60

Stiffeners Vertical plate plate

= 125 x 200 x 12 A4

25 200

= 300 x 700 x 14 A6

300

14

18

4,3

14

18

4,3

33°

,5 28 R2

125

20x20

° 23

100

700

,5 28 R2

50 150

23° 23 °

23 °

625 1250

50

1250

23°

475

40 R2 28 ,5

19 300 369

625 16

300

= 300 x 700 x 14 A6

300

= 290 x 1250 x 16 A 5A

230 290

121,5 700

16

121,5

Figure 7 – Details of an interior composite beam-to-column joint. (Dimensions in mm) In view of fire resistance, thermal analyses of joints were performed to obtain the internal temperature distributions. In this respect, a 3D finite model of an interior joint was implemented in the Abaqus software (Hibbitt et al., 2000) and employed to conduct thermal analyses for different fire exposures, i.e. 15, 30 and 60 min, respectively. The joint endowed with prefabricated slabs exhibited a better performance compared to the joint endowed with steel sheeting, see Figure 8, being the top concrete temperature one order of magnitude smaller.

a)

b)

Figure 8 - FE model and temperature distribution for an interior joint endowed with: a) prefabricated lattice girder slabs b) profiled steel sheeting slabs. Moreover, FE analyses allowed the component method to be applied to joints at different fire exposures. In detail, Figure 4(b) shows a considerable reduction of the joint moment strength as a function of fire exposure time. The hogging moment becomes approximately 20% of the initial value after 30 minutes of exposure and it approaches about 5% of its initial value after 60 minutes. Nonetheless, these performances are enough to evacuate the building after a quake.

TEST PROGRAM The experimental programme regarded the execution of six seismic tests and six fire tests on full-scale substructures representing interior welded beam-to-column joints. Seismic tests were carried out at the University of Trento, Italy, considering both cyclic and monotonic loadings. Fire tests were conducted at the Building Research Establishment, UK, with asymmetric loading on joints to simulate adjacent primary beams of different length. Table 1 reports the joint specimens subjected to cyclic and monotonic loading up to collapse. Specimens 2-5 were endowed with horizontal Nelson stud connectors welded around the column, see Figure 7, to increase the friction level between the concrete slab and the composite column, thus favouring the Mechanism 2 idealized in Figure 5b. Specimen were tested according to the ECCS stepwise increasing amplitude loading protocol, modified with the SAC procedure (ECCS 1986, SAC 1997). In detail, it was imposed a yield displacement ey= 0.005h = 17.5 mm, where h represents the storey height. Table 2 reports the test nomenclature of four interior specimens subjected to fire tests. While two specimens, i.e. T21 and T24 were pre-damaged to simulated damage owing to Type 1 spectrum compatible accelerograms at 0.4g pga (EN Eurocode 8-1, 2005), Test T22 and T25 were not, to clearly appreciate seismic damage effects on fire resistance. Table 1 – Specimens subjected to monotonic and cyclic tests N.

Label

Test Protocol

Type of Specimen

1

WJ-P1

Cyclic

Specimens with electro-welded lattice girder slabs and no Nelson connectors around the column

2

WJ-P2

Cyclic

Specimens with electro-welded lattice girder slabs and Nelson connectors around the column

3

WJ-PM

Monotonic

Specimens with electro-welded lattice girder slabs and no Nelson connectors around the column

4

WJ-S1

Cyclic

Specimens with profiled steel sheeting slab and no Nelson connectors around the column

5

WJ-S2

Cyclic

Specimens with profiled steel sheeting slab and Nelson connectors around the column

6

WJ-SM

Monotonic

Specimens with profiled Steel Sheeting slab and no Nelson connectors around the column

Table 2 - Specimens subjected to fire tests N.

Label

Test Method

Type of Specimen

1

T21

Fire

Pre-damaged Specimen endowed with steel sheeting slabs

2

T22

Fire

Undamaged Specimen endowed with steel sheeting slabs

3

T24

Fire

Pre-damaged Specimen endowed with prefabricated lattice slabs

4

T25

Fire

Undamaged Specimen endowed with prefabricated lattice slabs

To accurately simulate damage owing to seismic events, non-linear dynamic time histories were performed by using the IDARC-2D program (Valles et al., 1996). Experimental data of joints were used to define both hysteretic laws in IDARC-2D and damage domains according to the

Chai & Romstad criterion (Chai et al., 1995). The corresponding values of damage in joints provided by IDARC-2D simulations are gathered in Table 3 where average values for joints with prefabricated slab and steel sheeting slab are provided owing to the limited number of experimental data. The trend is evident: the damage in joints was limited and repairable. Table 3. Damage index for joints with prefabricated slabs and slabs with steel sheeting Joint typology Exterior joint Interior joint

Damage index D 0.43 0.34

Subsequently, specific deformations were prescribed on joints through monotonic vertical loading to induce the damage identified in Table 3. Hence, specimens were loaded according to the fire load combination (EN Eurocode 1991-1-2, 2004) and fire tests were undertaken. Typical beam-to-column joint specimens are illustrated in Figure 9. A welding procedure employed in laboratories was conceived to simulate on site welding, taking into account the Type of electrode employed, i.e. OERLIKON-ETC PH355 φ=3.25, with a pre-heating at 80°C. Slab with profiled steel sheetings

Slab with electro-welded lattice girders

A 150

A

CFT Ø 457 x 12

A

CFT Ø 457 x 12

IPE 400

IPE 400

A

a)

b)

Figure 9 - Beam-to-column joint specimens: a) slab with lattice girders; slab with sheeting The experimental set-up employed for seismic tests is shown in Figure 10. A hydraulic actuator endowed with a capacity of ±1000 kN and a stroke ±250 mm was employed. Different sensors were utilized: 5 inclinometers to measure rotations of joint and beams; 4 LVDTs to detect interface slip between steel beam and concrete slab; 2 LVDTs to measure joint deformations; 10 LVDTs and 4 Omega strain gauges to assess concrete slab deformations; 8 strain gauges to monitor axial deformations of rebars; 8 strain gauges to measure top and bottom plate deformations and flange strains; 2 load cell located in trusses to measure horizontal and vertical components of reaction forces; 1 digital transducer DT500 to detect top column displacements. SG STRAING GAUGE O OMEGA STRAING GAUGE LVDTs L

SLAB

SG L L L O SGL SG OO L L SG SG L L SG L O

I INCLINOMETERS L LVDTs

L

I

LL

I

I

I I LL

L

BOTTOM PLATE SG

SG

BEAM

SG SG

BEAM

SG SG

SG SG

TOP PLATE

Figure 10 - Lateral view of the test set-up and instrumentation

SG

TEST RESULTS AND ANALYSIS Seismic test results of beam-to-column joints Both the force-displacement and moment-rotation relationships of WJ-P1 and WJ-P2 specimens with electro-welded lattice slabs and without/with Nelson connectors around the column are illustrated in Figure 11 and 12, respectively. Plastic hinges developed in beams adjacent to joint and progressive deterioration of strength and stiffness was associated with beam flange buckling. Failure was due to beam flange cracking. A reader can observe that specimens exhibited a similar behaviour and beams developed plastic rotations greater than 25 mrad required by Eurocode 8-1 (2005) for moment resisting frames of Medium ductility class. Specimen WJ-P1

Specimen WJ-P1 1000

200

-14

-10

-6

-200 -2

2

6

10

14

M [kNmm]

Force [kN]

600

-80

-60

-40

-600 -1000

1000 800 600 400 200 0 -20-200 0 -400 -600 -800

20

40

80

Left Hinge

φ [rad]

Displacement [e/ey]

60

Right Hinge

Figure 11 – Specimen WJ-P1: Force-Displacement and Moment-Rotation relationship Specimen WJ-P2

Specimen WJ-P2 1000

200 -14

-10

-6

-200 -2

2

6

10

14

M [kNmm]

Force [kN]

600

-80

-60

-40

-600 -1000

1000 800 600 400 200 0 -20-200 0 -400 -600 -800

20

60

80

Left Hinge

φ [rad]

Displacement [e/ey]

40

Right Hinge

Figure 12 – Specimen WJ-P2: Force-Displacement and Moment-Rotation relationship Similar results were obtained for specimens WJ-S1 and WJ-S2 endowed with slabs with profiled steel sheeting and without/with Nelson connectors around the column, respectively. Experimental results are highlighted in Figure 13 and Figure 14, respectively. Favourable results can be observed. Nonetheless differently from previous cases, both in beams and joints, the neutral axis was located in the beam web for sagging moment owing to the greater damage imparted by lateral loads to the slabs endowed with steel sheeting. Specimen WJ-S1

Specimen WJ-S1 1000

-14

200 -10

-6

-200 -2

2

-600 -1000

Displacement [e/ey]

6

10

14

M [kNm])

Force [kN]

600

-80

-60

-40

1000 800 600 400 200 0 -20-200 0 -400 -600 -800

φ [rad]

20

40

60

80

Left Hinge Right Hinge

Figure 13 – Specimen WJ-S1: Force-Displacement and Moment-Rotation relationships

Specimen WJ-S2

Specimen WJ-S2 1000

200

-14

-10

-6

-200 -2

2

6

10

14

M [kNm]

Force [kN]

600

-80

-60

-40

-600 -1000

1000 800 600 400 200 0 -20-200 0 -400 -600 -800

20

40

80

Left Hinge Right Hinge

φ [rad]

Displacement [e/ey]

60

Figure 14 - Specimen WJ-S2: Force-Displacement and Moment-Rotation relationships Monotonic test results for the specimen WJ-PM endowed with prefabricated slabs with electrowelded lattice girders without Nelson connectors welded around the columns are plotted in Figure 15. Similar results were obtained for the specimen WJ-SM slab endowed with slabs with profiled steel sheeting (Bursi et al., 2008). The electro-welded lattice slabs of WJ-PM exhibited less damage owing a most favourable composite action in the plastic hinge beam section. Specimen WJ-PM

Specimen WJ-PM

Right Plastic Hinge 1000

1000

Left Plastic Hinge

M [kNm]

Force [kN]

800 600 400

600 200

-160

-120

-80

-40-200 0

40

80

120

160

200 -600 0 0

6

12

18

24

Displacement [e/ey]

-1000

30

φ [mrad]

Figure 15 - Specimen WJ-PM: Force-Displacement and Moment-Rotation curves Fire test results of beam-to-column joints Both pre-damaged and undamaged specimens were subjected to fire loading, see Table 2, and some results are presented herein. The Temperature vs. time curve imposed to the specimens T21-T22 and T24-T25 is shown in Figure 16(a) and (b), respectively. Specimens T21 and T22 with profiled steel sheeting slabs exhibited failure owing to an excessive rate of deflection at approximately 40 minutes. The test on specimen T21 terminated after approximately 34 minutes owing to runaway deflection. Following the fire test, the profiled steel sheeting separated from the slab; then the slab cracked both along the surface and through the depth with extensive buckling at one hour both of the lower flange and the web of the adjacent east beam, as shown in Figure 17. T24 and T25 specimens endowed with prefabricated slabs endured one hour of fire; however, in both cases specimens were very close to failure as indicated in Figure 16b, by an increasing rate of deflections towards the end of the test.

0

0 15

30

45

60

-25

-450 -900

Time [min]

T21 Temperature T22 Temperature T21 West Deflec. T22 West Deflec. T21 East Deflec. T22 East Deflec.

Temperature [°C]

25

Deflection [mm]

Temperature [°C]

50 450

T24 vs. T25

900

75

50

450 25

0

0

15

30

45

-900

a)

60

75

90

-25

-450

-50 -75

75

Time [min]

T24 Temperature T25 Temperature T24 West Deflec. T25 West Deflec. T24 East Deflec. T25 East Deflec.

Deflection [mm]

T21 vs. T22

900

-50 -75

Figure 16 – Performance of damaged (T21/T24) and undamaged (T22/T25) specimens.

b)

a)

b)

Figure 17 – Specimen T21: a) Surface cracking of the slab; b) Local buckling of east beam However, at this stage, there was no permanent deformation and no sign of any significant damage from fire tests. Hence, we can underline that: i) there was no noticeable difference in the fire performance between pre-damaged and undamaged specimens both with precast and steel sheeting slabs; this result is in agreement with damage values reported in Table 3 and with the inherent safety of composite joints (Eurocode 1998-1, 2005; Bursi et al., 2008); ii) precast slabs performed better also in fire tests than the corresponding specimens with steel sheeting at a fire exposure in excess of the 15 minutes required; iii) all specimens inherited favourable seismic properties by performing in a ductile manner also under fire loading. Analysis results of moment resisting frames At the frame level, several simulations were performed to assess both the seismic and fire behaviour of moment resisting frames endowed with the proposed joint both with steel sheeting and prefabricated slabs (Bursi et al., 2008). As observed above, specimens subjected to cyclic loadings developed plastic rotations greater than 25 mrad, thus being adequate for moment resisting frames of Medium ductility class (EN Eurocode 1998-1, 2005). Along the line of Aribert et al (2006) work, incremental dynamic analyses were performed on the prototype frames depicted in Figure 1, in order to find any correlation between required plastic rotations for high ductility moment frames, i.e. about 35 mrad, interstorey drifts and p.g.a. We found that by using artificial accelerograms compatible with Type 1 Eurocode 8 spectrum (2005), interstory drifts between 4.09÷6.78 per cent developed with p.g.a. ranging between 1.5÷2.5g. Moreover, it was shown that energy dissipating mechanisms within frames endowed with examined joints relied on the formation of plastic hinges at beam ends and that a behaviour factor of about 4 was estimated, thus allowing to adopt the proposed joints for Ductility Class M structures. Several simulations were carried out on the frames depicted in Figure 1 to assess the fire performance of proposed joints in moment resisting frames endowed with composite slabs. The effect of the seismic loading applied prior to fire loading was taken into account by imposing one loading-unloading cycle through identical horizontal forces applied at each floor. In addition to an initial imperfection, this loading cycle induced some plasticity in each frame. Figure 18 shows the relationships between horizontal loads and residual drifts after frame unloading at ambient temperature. The impact of the earthquake on the fire resistance of the analysed frames appeared to be not so significant because failure occurred when a beam plastic mechanism formed in the long-span heated beams. Moreover, no global instability modes formed under fire loading with a fire exposure lower than 30 minutes in the majority of cases.

0.040 0.035 0.030

δ 0.025

H

0.020 0.015 0.010

Steel sheeting slab

Lattice girder slab

0.005 0.000 0

50

100

150

200

250

300

Figure 18 - Evolution of the residual rotation δ/H (mrad) as a function of the horizontal load (kN) CONCLUSIONS A multi-objective design methodology dealing with seismic-induced fire actions on steelconcrete composite moment resisting frames endowed with full strength joints using concrete filled tubes has been presented in this paper. Instead of a traditional single-objective design where fire safety and seismic safety are achieved independently and the sequence of seismic and fire loadings are not accounted for, in the multi-objective design developed by means of mechanical resistance of members and joints and proper details it has been guaranteed: i) seismic safety with regard to accidental actions; ii) fire safety for at least 15 min fire exposure on a structure characterised by stiffness deterioration and strength degradation owing to seismic actions. As a result, the fire design applied to a structure with reduced capacity owing to seismic actions has achieved structural, seismic and fire safety as required; but also structural and fire safety and structural and seismic safety, respectively. The goals have been achieved through a balanced combination of analytical/numerical and experimental work. ACKNOWLEDGMENTS The writers are grateful to the European Union for financial support under the project PRECIOUS-RFS-CR-03034 and to the partners as well. Nonetheless, conclusions of this paper are those of the authors and not of the sponsors agency. REFERENCES Aribert J.M., Ciutina A.L., Dubina D., 2006, “Seismic response of composite structures including actual behaviour of beam-to- column joints”, in Composite construction in steel and concrete V, Leon and Lange ed. South Africa. ASCE, 708-717. Block F.M., Burgess I.W., Davison J. B., Plank R. J. 2007 The development of a componentbased connection element for endplate connections in fire. Fire Safety Journal 42. 498–506. Bursi O.S., et al. Editors, Final Report PRECIOUS Project Contr. N. RFSCR-03034, 2008. Prefabricated Composite Beam-to-Concrete Filled Tube or Partially Reinforced-ConcreteEncased Column Connections for Severe Seismic and Fire Loadings. Chai Y.H., Romstad K.M., Bird S.M. 1995, Energy-based linear damage model for high-intensity seismic loading”, Journal of Structural Engineering, ASCE, 121(5), 857-864.

ECCS. 1986. Recommended Testing Procedure for Assessing the Behaviour of Structural Steel Elements under Cyclic Loads. Publication n° 45; EN 1991-1-2. 2004. “Eurocode 1: Actions on Structures. Part 1-2 : General Actions – Actions on structures exposed to fire”; EN 1993-1-8. 2005. “Eurocode 3: Design of steel structures - Part 1-8: Design of joints”; EN 1993-1-2. 2005. “Eurocode 3: Design of steel structures. Part 1-2: General rules - Structural fire design”; EN 1994-1-2. 2005. “Eurocode 4: Design of composite steel and concrete structures - Part 1.2: General rules – Structural fire design”; EN 1998-1. 2005. “Eurocode 8: Design of structures for earthquake resistance - Part 1: General rules, seismic actions and rules for buildings”; Franssen J.-M. 2000. “SAFIR; Non linear software for fire design”. Univ. of Liege; Gil B., Bayo E. 2008. An alternative design for internal and external semi-rigid composite joints. Part I: Experimental research Engineering Structures 30. 218–231 Gil B., Bayo E. 2008. An alternative design for internal and external semi-rigid composite joints. Part II: Finite element modeling and analytical studies. Engineering Structures 30. 232–246 Hibbitt, Karlsson and Sorensen 2000. “ABAQUS User’s manuals”. 1080 Main Street, Pawtucket, R.I. 02860; Kobe City Fire Department 1995 Investigation report on damages fire protection systems caused by the Hanshin-Awaji earthquake in Kobe. SAC 1997, Protocol for Fabrication, Inspection, Testing, and Documentation of Beam-Column Connection Tests and Other Experimental Specimens, report n. SAC /BD-97/02; Salvatore W, Bursi OS, Lucchesi D. 2005. Design, testing and analysis of high ductile partialstrength steel–concrete composite beam-to-column joints. Computers & Structures; 83 (28– 30):2334–52. Santiago, A, Simões da Silva L, Vila Real P, Milan Veljkovic. 2007. Numerical study of a steel sub-frame in fire. Computers and Structures. (submitted). SEAOC - Structural Engineers Association of California. 1995. Performance Based Seismic Engineering of Buildings, Vision 2000 Committee, 2 vols. Sacramento, CA. Simões da Silva, L., Santiago, A., Moore, D. e Vila Real, P. 2005. Behaviour of steel joints and fire loading. Int. J. of Steel and Composite Structures 5(6), pp. 485-513 (2005). The Marine and Fire Insurance Association of Japan (MFIAJ) 1995. Study report on reliability of fire protection systems at an earthquake, 1995. Valles R.E., Reinhorn A.M., Kunnath S.K., Li C., Madan A. 1996, IDARC2D Version 4.0: A Program for the Inelastic Damage Analysis of Buildings, Tech. Report NCEER-96-0010, 1-8. Yashiro, Y., Ebihara, M. and Notake, H., 2000, Fire safety Design and Fire Risk Analysis Incorporating Staff Response in Consideration of Fire Progress Stage. 15th meeting of USJapan Cooperative Program in Natural Resources by Int. Assoc. for Fire Safety Science.