The interaction between the bending moment and axial loads in the heated internal joint after the loss of the column is the objective of the tests. The effect of the localised fire (that led to the column loss) was simulated by the application of elevated temperatures in the composite joint zone. The temperatures considered in these tests were based on previous observations: in the past, real tests performed in composite steel-concrete open car park buildings subject to burning cars showed that the beam bottom flanges temperature was lower than 500ºC (Jaspart et al., 2008); however temperatures of 700ºC were measured in recent tests performed in France (Joyeux et al., 2002), probably due to the manufacture evolution of cars, with more combustible plastic materials as well as higher petrol tank capacity. Seven beam-to- column sub-frames were tested in the University of Coimbra: one reference test at ambient temperature; five tests at 500ºC or 700ºC; and a demonstration test, for which the sub-frame was subject to an increase of the temperature up to the failure of the column. The effect of the axial restraint to beam coming from the unaffected part of the building was also studied: three tests without beam axial restraint; two tests with total beam axial restraint; and two tests with realistic beam axial restraint (Figure 1).
T1 (20ºC) T2 (500ºC) T3 (700ºC) T4 (500ºC) T5 (700ºC) T6 (700ºC) T7 (Fire - Dem.) Figure 1. Seven experimental tests
Connections in Steel Structures VII / Timisoara, Romania / May 30 - June 2, 2012 375 376 Connections in Steel Structures VII / Timisoara, Romania / May 30 - June 2, 2012
2.2. Sub-frame and testing arrangement
A typical open car park structure was specially designed for the European ROBUSTFIRE project (Demonceau et al., 2012). This building was defined as the most common as possible in order to obtain, at the end of the project, general design rules for such structures, ensuring sufficient robustness under fire conditions. The selected structure was a braced open car park building with eight floors of 3 m height, composite slabs, composite beams and steel columns. The tested sub-frame was selected from the fifth floor of this real car park building; however, because of the laboratory dimensions, the beam length was reduced from 10 m in the real building to 3 m in the tested sub-frame. The sub-frame was defined by two unprotected composite beams IPE 550 steel cross-sections, grade S355, and one unprotected HEB 300 cross-section steel column, grade S460 (Figure 2). A reaction frame, perpendicular to the plane of the sub-frame, was used to fix the hydraulic jack, which was linked by a pin to the top of the column. The column base was hinged and fixed to a reinforced concrete footing. The hydraulic cylinder located at the column base (except for the last test 7) simulated the progressive loss of the column (by decreasing the oil pressure). A smaller steel profile HEB 140 cross-section was used at the bottom column in order to allow: i) the concreting of the specimens (the composite sub-frames were located at the floor level), ii) the simulation of the column loss under fire in test 7 (buckling failure). The column was restrained at the top of the joint and at the bottom column (column restraint in Figure 2). This restraints system avoided any horizontal displacement or rotation in the plane or out of the plane of the sub-frame.
Cylinder HEB 300 IPE 550 ST RONG W A LL HEB 1 40 HEB 2 2 0 HEB 140 HEB 2 2 0 HEB 1 40 Horizontal beam support HEB 1 40 Bottom column (Test 7) Concrete footing Hydraulic Jack COMPOSITE SLAB AND COMPOSITE BEAM
Beam vertical support 3 m 2.2 m 2.5 m 4 m 3.8 m Beam vertical support Reaction frame Bottom column
Steel footing Steel footing
HEB 300 HEB 300 Column restraint Column restraint Column restraint 14.1 m Load cell TOTAL AXIAL RESTRAINT TO THE BEAM REALISTIC AXIAL RESTRAINT TO THE BEAM (Spring) NO AXIAL RESTRAINT TO THE BEAM HEB 300 HEB 300
Left side Right side
Figure 2. General layout, longitudinal view
The steel beam was fully connected to the 130 mm thickness composite slab by 22 shear studs (diameter = 19 mm; height = 100 mm). The composite joint, with the bolts M30 grade 10.9 is shown in Figure 3. In order to ensure the composite behaviour of the beam-to-column connections, ten steel rebars of diameter 12 mm were placed in the composite slab (five at each side of the column).
5 φ12 130 IPE 550 210 80 260 80 590 HEB 300 900
Figure 3. Tested joint 2.3. Description of the loading sequence
The tests 1 to 6 were performed under constant temperature and they were divided into 3 main steps: step 1 - application of an initial hogging bending moment in the joint as in the real car park (-450 kNm in test 1 at ambient temperature, and - 236 kNm in tests 2 to 6, as defined in Demonceau et al. (2012); step 2 - heating of the joint zone with a linear rate of 300ºC/hour (up to reach 500ºC or 700ºC in beams bottom flanges); and step 3 - simulation of the progressive loss of the column and increase of the joint sagging bending moment (by increasing the vertical load at the column top) up to the failure of the joint under constant temperature.
The test 7 (demonstration test) simulated the loss of the column due to the increase of temperatures under constant loading. This test was divided into 4 steps:
step 1 - application of an initial hogging bending moment in the joint (-236 kNm);
step 2 - application of a constant compression load at the column top (+250 kN).
Due to the capacities of the loading and measuring equipment, the steel section of the bottom column was reduced from HEB 300 to HEB 140 in order to reach the column buckling failure under 800ºC or less when it is subject to the 250 kN at the column top; step 3 - heating of the joint zone and the bottom column, respectively up to 400ºC (measured in the beam bottom flanges) and 800ºC (the column should failed under lower temperatures), under faster heating rate in the bottom column. The joint temperature was limited to 400ºC in order to avoid the joint failure before the collapse of the column; this value was defined taking into account the constant temperature tests: at 500ºC, the load at the column top reached maximum 400 kN in test 2, and under 700ºC, it reached 200 kN; moreover, additional compression load due to thermal expansion effects should be considered; step 4 - heating of the joint zone up to the failure of the sub-frame (the load at the column top (+250 kN) and the temperature in the bottom column (800ºC) were kept constant).
2.4. Mechanical and thermal loadings
Steel temperatures in tests 2 to 7 were increased using Flexible Ceramic Pad heating elements (concrete was not heated). In tests 2 to 6, the heated zone was defined by a length of 0.6 m of the beam to each side of the joint, the bolts and 1 m of column. In test 7 (demonstration test), the bottom column (HEB 140) was heated, and the joint zone was reduced to a beam length of 0.4 m. Servosis hydraulic jack (Fmax. = 1000 kN; ∆max. = 280 mm) was used to apply the mechanical loading at the
column top. Three different restraints stiffness’s from the unaffected part of the building
IPE 550
HEB 30
0
Connections in Steel Structures VII / Timisoara, Romania / May 30 - June 2, 2012 377 378 Connections in Steel Structures VII / Timisoara, Romania / May 30 - June 2, 2012
were considered: tests 1, 2 and 3 - no beam axial restraint; tests 4 and 5 - total beam axial restraint; and tests 6 and 7 - realistic beam axial restraint (spring restraint). When no restraint was applied, the beams were free to the axial movement. For the total beam axial restraint, a steel beam HEB 300 cross-section, linked from the end of each tested beam to strong walls, was used to totally restrain the beam in the axial direction, allowing the rotation. The spring restraints were simulated using hydraulic cylinders, each one being independently and manually controlled in order to adjust the spring stiffness (50 kN/mm).