In order to simplify the presentation of tests results as well as the comparisons between them, the results of only one connection from each tested internal joint are presented:leftconnectionsfortests1,2,3,5and7,andrightconnectionfortests4and6. 4.1. Temperature results
Tests2to6wereperformedunderconstanttemperatures,controlledinthebeams bottom flanges (at 250 mm from each end-plate). Figure 5a depicts the temperatures
evolution during test 3 (700ºC) in the left side of the frame: in the beam, at 200 mm from the connection (in the bottom flange, web and top flange), in the column flange and web, in bolt from row 4, and in the concrete rib in contact with the steel beam near the joint. The reduced web thickness allowed a faster temperature increase; because top flanges were shielding by the concrete slab and heated only by heat transfer, steel temperatures were much lower than the one measured in the lower flanges and webs. In all the tests, concrete temperatures did not rise above 300ºC. More information about the temperature results for each test can be found in Demonceau et al. (2012).
Figure 5. Evolution of the temperatures during a) test 3 at 700ºC, b) test 7 (demonstration test)
Figure5b presents the temperatures evolution during the entire test 7. In step 3 (1st increase of temperatures), the beams bottom flanges were heated up to 400ºC
(rate of 200ºC/hour), whereas webs, joint components and column HEB 300 reached slight lower temperatures (between 300ºC and 400ºC). The bottom column (HEB 140) reached its maximum resistance capacity and failed under the average temperature of 658ºC (686ºC measured in flange at the centre of the column). In step 4 (2nd increase of the temperatures), the temperatures in the beams bottom flanges attained 800ºC (rate of 300ºC/hour), and finally, the entire sub-frame collapsed. The maximum temperature measured in the shear studs in the composite slab, near the joint, was 179ºC, and temperature of 200ºC was measured in the concrete slab.
4.2. Mechanical results and failure modes of tests 1 to 6
Figure 6 depicts, for each test, the evolution of the bending moment versus the rotation of each connection. The joint rotation was estimated using the vertical displacements measured at: i) 1500 mm from the end-plate, and ii) at the column top. The bending moment was calculated using the reaction loads at the beams supports and ends. In step 1, the applied hogging bending moment reached: -501 kNm for the reference test 1, and around -236 kNm in tests 2 to 6. At the beginning of the heating phase (step 2), the reaction loads increased under thermal expansion effects and they reached a minimum value: the minimum hogging bending moments attained around -500 kNm in tests 3, 5 and 6 (under 700ºC), and around -357 kNm
Connections in Steel Structures VII / Timisoara, Romania / May 30 - June 2, 2012 379 380 Connections in Steel Structures VII / Timisoara, Romania / May 30 - June 2, 2012
in tests 2 and 4 (under 500ºC). Finally, these reaction loads decreased because of the degradation of steel properties at elevated temperatures.
Under sagging bending moment (step 3), concrete crushing in compression was the first failure mode, due to the joint rotation and the resulting high compressive strain at the upper concrete surface: first the concrete crushed against the column flanges, then the entire slab width failed at the upper concrete surface, and finally over the entire thickness. Bolts failures happened in tests 1, 2 and 6 (respectively under 48 mrad, 74 mrad and 83 mrad of connection rotation for the first bolt failure); the other tests were ended because the maximum vertical displacement of the hydraulic jack at the column top was reached. The failed bolts were identified in the bottom bolt rows, because of higher tensile forces under sagging bending moment.
Figure 6. Joint bending moment versus rotation at the connection
The evolution of the bending moment at the joint versus the axial restraint load is presented in Figure 7. The restraints were connected to the beams since the beginning of the test. During step 1, the reaction loads and displacements due to the mechanical loadings were not sufficient to develop axial forces in the beams axial restraints. During the heating phase (step 2), due to the thermal expansions of beams, the beams ends moved in the outward direction, and compression loads were applied by the restraints. After the column loss (step 3), the axial restraints increased the compression loads because beams ends continued moving in the outward direction. Two reasons are highlighted: i) when the column began to go downward, the joint was subject to sagging bending moment: the concrete slab was in compression against the column flange whereas the joint bottom part was in tension. While the concrete was resisting, it prevented any inward horizontal displacement of the composite section. It was only after the concrete crushing in compression (or after large vertical column displacement) that the composite beam began to displace inward; ii) the horizontal displacement of the beam end was measured at the centroid of the steel beam but not at the centroid of the composite section, so, when the beam rotated, the measured displacements at the centroid of the steel beams ends continued to show an outward movement even after inward movement of the centroid of the composite beams.
In tests 4 and 5 (total axial restraint), the concrete crushed under smaller joint rotations (around 25 mrad) in comparison to test 6 (35 mrad) due to the much higher axial restraint loads developed when the beams ends tried to displace outward while they were restrained. Finally, tensile loads were observed at the end of the test 6.
Figure 7. Joint bending moment versus axial load restraint
Figure 8 shows the final deformations of test 1 (20ºC) and test 3 (700ºC). The steel end-plates deformed in the bottom and centre parts and showed a high ductility of this joint. The deformation at the centre of the end-plate happened because of the joint configuration: i) 4 bolt rows and a high distance between the rows 2 and 3 (260 mm), ii) the end-plate (15 mm) was thinner than the column flange (19 mm), and iii) an initial gap was noticed just after the bolts pre-loading (0.6 mm was measured for the reference test). Moreover, it seems that the beam web was pulling the end- plate due to the sagging bending moment (tensile loads at the bottom part). Due to high stresses/deformations, a crack at the base steel end-plate, just above the weld, was observed at the end of the test 1 at ambient temperature (Figure 8a).
Figure 8: Deformations of the connections in: a) test 1 (ambient temperature), and b) test 3 (700ºC)
Connections in Steel Structures VII / Timisoara, Romania / May 30 - June 2, 2012 381 382 Connections in Steel Structures VII / Timisoara, Romania / May 30 - June 2, 2012
Table 1 summarizes the failure modes and the local deformations of each test. Table 1. Failure modes and local deformations of each test
Test Temp
(ºC)
Restraint (kN/mm)
T1 20 0
Concrete crushing in compression; failure of 2 bolts in tension (left side); crack at the end-plate (bottom – left).
Local deformations at the end-plates (bottom and centre). T2 500 0 Concrete crushing in compression; failure of 3 bolts in tension (left side)
Local deformations at the end-plates (bottom and centre).
T3 700 0
Concrete crushing in compression; failure of 2 bolts in tension (left side) during the cooling phase.
Local deformations at: the end-plates (bottom and centre); column web (bottom part); beam bottom flange (left). T4 500 Total
Concrete crushing in compression; failure of 2 bolts in tension (right side) during the cooling phase.
Local deformations at: the end-plates (bottom and centre); column web (top part); beams webs; column flange (left). T5 700 Total
Concrete crushing in compression.
Local deformations at: the end-plates (bottom and centre); column web (bottom part); beam bottom and top flanges (right).
T6 700 50
Concrete crushing in compression; failure of 3 bolts in tension (2 on the right - 1 on the left).
Local deformations at: the end-plates (bottom and centre); beams bottom flanges.
T7* 400; 800 50
Bottom column failure; concrete crushing in compression; failure of 3 bolts in tension (left side); crack at the end-plate (bottom - left).
Local deformations at the end-plates (bottom and centre) * Test 7 (demonstration test) is presented in §3.3.
From the results of tests 1 to 5, the effect of the temperatures and the effect of the axial restraint to the beam on the joint behaviour can be discussed. Table 2 summarises the values of the maximum sagging bending moment reached for each test, as well as the corresponding axial load and connection rotation.
Table 2. Comparisons of the results related to maximum sagging bending moment M+max (tests 1 to 6)
T1 T2 T3 T4 T5 T6
Temperature 20ºC 500ºC 700ºC 500ºC 700ºC 700ºC
Beam axial restraint 0 0 0 total total 50 kN/mm
M+max(kNm) 710.1 565.0 357.1 746.4 828.0 355.5
Rotation θM+max (mrad) 46.9 49.5 92.3 54.9 43.0 55.1
Axial load (kN) 0 0 0 -990.7 -1646.7 -297.3
4.2.1. Effects of the temperature
The three first tests, performed without beam axial restraint, showed that the increase of temperature:
• Decreases the maximum sagging bending moment of the joint (by 20% between test 1 (20ºC) and test 2 (500ºC) and by 50% between test 1 and test 3 (700ºC);
• Increases the capacity of rotation and the ductility of the joint, by i) increasing the connection rotation corresponding to the maximum sagging moment (by 97% from test 1 to test 3, and by 87% from test 2 to test 3); ii) increasing the maximum joint rotation at the first bolt failure (in tests 1 and 2, the first bolt failed for respectively 49 mrad and 74 mrad of rotation (increase of 55%), and no bolts were failed at 132 mrad of rotation in test 3);
• Decreases the initial stiffness, by 36% and 49% between test 1 (20ºC) and respectively test 2 (500ºC) and test 3 (700ºC).
In test 1, the maximum sagging bending moment was equal to 710 kNm, which is 21% higher than the theoretical value (588 kNm) calculated in Demonceau et al. (2012).
4.2.2. Effect of the axial restraints to the beams
Tests 2 and 4 were both performed under 500ºC, respectively with none and total beam axial restraints, and tests 3, 5 and 6 were performed under 700ºC, respectively with none, total and spring beam axial restraints. It was observed that the beam axial restraint, and consequently, the compression axial load in the joint:
• Increases the maximum sagging bending moment, by 32% between test 2 and test 4, under 74 mrad of joint rotation (corresponding to the first bolt failure in test 2), due to the axial compression load equal to 773 kN in test 4. Note that between tests 3 and 6, the maximum bending moment was not affected by the axial restraint to the beam (difference of 0.5%); however, the corresponding rotation was 40% lower in test 6;
• Increases the joint rotation capacity and, consequently, the joint ductility: the compression load from the axial restraint combined with sagging bending moment, moved the neutral axis of the connection downward, allowing the development of additional compression loads in the concrete slab, and the reduction of the tensile loads in the bottom bolt rows. Once the concrete crushed against the column slab and along the entire slab width, tests 4 and 5 were still able to continue to deform without failure: between the maximum sagging bending moment and the end of the test, the rotation increased by 113% in test 4 and by 184% in test 5;
• Increases the initial stiffness of the bending moment/rotation curves, by 83% at 500ºC (from test 2 to test 4) and by 54% at 700ºC (from test 3 to test 5). 4.3. Mechanical results and failure modes of test 7 (demonstration test)
The bending moment/rotation and bending moment/axial load curves were depicted in Figure 6 and Figure 7. The hogging bending moment was initially reached during step 1 (-281 kNm). During step 2, the hydraulic jack increased the load at the column top up to reach +250 kN. The axial force in the bottom column was then equal to 341 kN (due to the additional reaction loads created by the initial hogging bending moment applied in the joint). In step 3, the reaction loads first increased
Connections in Steel Structures VII / Timisoara, Romania / May 30 - June 2, 2012 383 384 Connections in Steel Structures VII / Timisoara, Romania / May 30 - June 2, 2012
under thermal expansion effects, and the axial force in the bottom column reached 604 kN (hogging bending moment equal to -505 kNm); the bottom column reached its maximum resistance capacity under 658ºC and failed. The failure of the column was progressive, and was defined as the moment at which the vertical reaction load came back to its initial value at the beginning of the step 3 (341 kN). As in tests 4, 5 and 6, compression axial restraint loads were developed during step 3, and continued to increase in step 4. During this step 4, the temperature in the joint increased under the constant load (+250 kN) applied at the top of the column and reached 770ºC in the beam bottom flanges (see Figure 5b). At this moment, the concrete slab was crushed in compression against the column flanges, and the vertical displacement increased faster. Once the concrete slab was crushed along the entire slab width, the sub-frame rapidly collapsed (beam bottom flanges temperature reached 800ºC). For security reasons, the test was stopped at a vertical displacement equal to 280 mm and left connection reached 150 mrad of rotation. The day after the test showed the failure of three bolts in the bottom bolt rows of the left connection. However, bolts failures were not observed on bending moment/rotation curve because failures were very smooth due to the increase of ductility of the bolt after 500ºC (see Figure 4). The steel end-plates also deformed in the bottom and centre part and due to high stresses/deformations, a crack at the base steel end-plate, just above the weld, was also observed (Demonceau et al., 2012).
5. CONCLUSIONS
The main objective of the experimental tests was to observe the combined bending moment and axial loads in the heated composite steel-concrete joint after the loss of the column. The effect of the localised fire (that led to the column failure) was simulated by the application of elevated temperatures in the composite joint zone. Seven beam-to-column sub-frames were tested: 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 studied.
In the six first tests, a hogging bending moment was initially applied in the joint, followed by a variation of this moment during the increase of temperatures. Then the column loss was simulated under constant temperature, and the sagging bending moment was increased. The first failure observed in all the tests was the concrete crushing in compression; some bolts from the bottom bolt rows failed later in tension under higher joint rotations in tests 1, 2 and 6; similar local deformations at the centre and bottom part of the end-plate were observed in all the tests. From tests performed without axial beam restraint, it was observed that the joint rotation capacity, as well as the ductility, was increased with the temperature, whereas the maximum sagging bending moment was lower. The advantage of the compression axial loads coming from the beams restraints was the increase of: i) the joint rotation capacity and its ductility, and ii) the capacity of the joint to sustain a higher sagging bending moment without bolt failure: the compression load from the axial restraint combined with sagging bending moment, moved downward the neutral axis of the connection, allowing the development of additional compression loads in the concrete slab, and reducing tensile loads in the bottom bolt rows.
The objective of the demonstration test was to show the real behaviour of the sub-frame when subject to a localised fire which leads to the loss of the column. A constant gravity load was applied to the frame in order to approximate to the real behaviour of the structure, and then beams, joint and column were heated in order to reproduce a localised fire (the temperature in the bottom column has raised faster than the temperature in the beam and joint). The bottom column failed under 658ºC; after that, it was observed that once the beams reached 700ºC, the vertical displacement of the joint under sagging bending moment began to increase faster, then the composite slab crushed, and finally the sub-frame collapsed. In this demonstration test, the sub-frame shown sufficient robustness provision; the loss of the column did not led directly to fragile failure of the joint, because of the ductility and capacity of the joint to deform and continue transferring loads. Besides, the composite slab considerably increased the joint resistance under the localised fire.
ACKNOWLEDGMENTS
The research leading to these results has received funding from the European Community's Research Fund for Coal and Steel (RFCS) under grant agreement n° RFSR-CT-2008-00036. Material support provided by PECOL, FELIZ and MARTIFER are gratefully acknowledged.
REFERENCES
[1] Abaqus (2010), Theory Manual & Users Manuals, Version 6.10, Hibbitt, Karlsson and Sorensen, Inc. USA.
[2] Akbarpour, M.R. and Ekrami, A. (2008). “Effect of temperature on flow and work hardening behavior of high bainite dual phase (HBDP) steels”, Materials Science
and Engineering, 475(1-2), pp. 293–298, 2008.
[3] Demonceau J.F., Huvelle C., Comeliau L., Hoang L. V., Jaspart J.P., Fang C., et al. (2012) “Robustness of car parks against localised fire”, European Commission, Grant Agreement Number RFSR-CT-2008-00036, Final report.
[4] EN1991-1-7:2002,“Eurocode1-Actionsonstructures-Part1-7:Generalactions- Accidental actions”, European committee for standardization, July 2006.
[5] Jaspart, J.-P.and Demonceau, J.-F. (2007). “Contribution to the derivation of robustness requirements for steel and composite structures”, 5th International Conference on Advances in Steel Structures, Editors: J.Y. Richard Liew & Y.S. Choo, Singapore, 5-7 December.
[6] Jaspart J.-P., Demonceau J.-F., Fang C., Izzuddin B., Elghazouli A., Nethercot D., et al. (2008) “Deliverable I: Definition of the problem and selection of the appropriate investigation ways”, Robustness of car parks against localised fire, Grant Agreement Number RFSR-CT-2008-00036.
[7] Joyeux, D., Kruppa J., Cajot L.G., Schleich J.B., van de Leur P. and Twilt L. (2002). “Demonstration of real fire tests in car parks and high buildings”. European Commission, Contract no 7215-PP/025, CTICM, Final Report, EUR 20466EN. [8] Le Pense, “Les parkings aériens en structure acier: développements récents /
Multi-storeycarparksinsteelstructure:lastdevelopments”,LaRevuedeMétallurgie - CIT, p867-874, Octobre 2002.
Connections in Steel Structures VII / Timisoara, Romania / May 30 - June 2, 2012 385 386 Connections in Steel Structures VII / Timisoara, Romania / May 30 - June 2, 2012