- Email: [email protected]
Behaviour of continuous concrete filled steel tubular columns loaded concentrically in fire
Journal of Constructional Steel Research 136 (2017) 101–109
Contents lists available at ScienceDirect
Journal of Constructional Steel Research journal homepage: www.elsevier.com/locate/jcsr
Behaviour of continuous concrete filled steel tubular columns loaded concentrically in fire
MARK
Kingsley U. Ukanwaa, Umesh Sharmab, Stephen J. Hicksc, Anthony Abud, James B.P. Lima, G. Charles Cliftona,⁎ a
Department of Civil and Environmental Engineering, The University of Auckland, Auckland, New Zealand Department of Civil Engineering, India Institute of Technology, Roorkee, India c New Zealand Heavy Engineering Research Association, HERA House, Auckland, New Zealand d Department of Civil and Natural Resources Engineering, University of Canterbury, Christchurch, New Zealand b
A R T I C L E I N F O
A B S T R A C T
Keywords: Fire resistance Concrete-filled steel tubular columns Continuous columns Fibre reinforced concrete
Design recommendations for concrete filled steel tubular (CFST) columns in fire are based on the results of experimental standard fire testing of CFST members where the same temperature is applied to the column over the full column height. However, this is not representative of a CFST column in a typical building, which is continuous between floors and which, in fire, is subjected to severe fire conditions on one floor at a time while the floors above and below remain cooler. In the experimental tests described in this paper, the columns are of 3.2 m height with the fire applied only to the central 2 m. Significant differences are observed between these tests and those previously conducted due to the partial length heating. In total, ten tests are conducted; the tests cover three different types of infill: plain concrete; bar reinforced concrete; and steel fibre reinforced concrete. End restraint conditions of fixed-fixed (F-F) and pinned-fixed (P-F) are considered; the axial load levels are between 0.33 to 0.38 of the squash load. The longitudinal elongation of the steel tube was less than 3 mm. Using the experimentally measured structural fire resistance (R), the axial capacity in fire was calculated in accordance with the codes of practice and are Compared with the experimentally tested structural fire resistance, showing that in some instances current design practice can be un-conservative.
1. Introduction Unprotected concrete filled steel tubes (CFST) (i.e. with no insulation material applied to the outside of the steel tube) can have an inherent fire resistance, due to concrete preventing inward buckling of the steel tube and hence increasing its local and global buckling resistance and the steel tube partially or totally confining the concrete, to give the column a higher compressive strength. In addition, the steel tube acts as a partial radiation shield to the concrete [1]. In the literature, the performance of CFST under ISO 834 [2] Standard Fire conditions has been considered by, amongst others, Romero et al. [3], Kodur and Latour [4] and Lie and Irwin [5], and has covered three types of in-fill; plain concrete, steel fibre reinforced concrete and rebar reinforced concrete. Finite element analyses studies have also been reported [6–9] which cover CFST columns filled with normal strength and high strength concrete loaded axially with high utilization factor in fire. The finite element analysis studies, however, have not covered continuous columns in fire.
⁎
Corresponding author. E-mail address: [email protected] (G.C. Clifton).
http://dx.doi.org/10.1016/j.jcsr.2017.05.011 Received 16 January 2017; Received in revised form 10 May 2017; Accepted 12 May 2017 0143-974X/ © 2017 Elsevier Ltd. All rights reserved.
In all previously reported experimental tests and finite element analyses, the entire column height is heated. In New Zealand, these columns are required to be continuous over their full length for seismic design requirements. The full length heating of the column is not representative of a fire cell in a continuous column construction where the fire is on one floor and the floors above and below remain cooler. This paper considers the behaviour of a continuous column heated over part of its length at one time. The experimental tests reported herein consider the issue of partially heated length in a continuous column and different concrete infills having measured compressive strength ranging from 80 MPa to 95 MPa. In total ten tests were conducted, with temperature only applied to 2 m of a 3.2 m high column (see Fig. 1); all tests were conducted on square hollow steel sections having nominal crosssections dimensions of 200 mm × 200 mm × 6 mm and 220 mm × 220 mm × 6 mm; the yield strengths of the sections tested ranged from 461 MPa to 569 MPa. The tests cover three different types of infill: plain concrete; bar reinforced concrete; and steel fibre
Journal of Constructional Steel Research 136 (2017) 101–109
K.U. Ukanwa et al.
Nomenclature As,T Ac,T Ar,T Am/V Es,T Ec,T Eƒi,Exp Eƒi,Rd Ec,sec,T Er,T (EI)fi F-F F-P ƒc ƒr ƒs
Ic,T Is,T
cross-sectional area of steel profile at the temperature °c cross-sectional area of concrete at the temperature °c cross-sectional area of reinforcement at the temperature °c section factor modulus of elasticity of steel tangent modulus of concrete at the temperature °c design effect of actions in fire situation for laboratory experiment design effect of actions in fire situation for design code secant modulus of concrete at the temperature °c modulus of elasticity of reinforcement at the temperature °c effective flexural stiffness in fire situation Fixed - Fixed Fixed - Pinned compressive cylinder strength of concrete at room temperature yield strength of reinforcement at room temperature yield strength of structural steel at room temperature
Ir,T k Le,T Nc,ƒi,Rd Nb,ƒi,Rd.,t R t αc T Tc Ts Tr λr λ,T ηƒi δ
reinforced concrete [10–13]. End restraint conditions of fixed-fixed (FF) and pinned-fixed (P-F) were considered. Compressive load levels ηfi of 0.33 and 0.38 were applied, with the loads determined in accordance
second moment of area of concrete at the temperature °c second moment of area of steel profile at the temperature °c second moment of area of reinforcement at the temperature °c effective length factor buckling length of column in fire situation design cross-sectional plastic resistance to axial compression in fire design axial buckling load of column in fire situation structural fire resistance steel tube thickness member slenderness reduction factor temperature temperature of concrete temperature of steel temperature of reinforcement relative slenderness of column at room temperature relative slenderness of column in fire situation design load level in fire condition steel contribution ratio
with the provisions in DR AS/NZS 2327 [14], which are based on the recommendations of Espinos et al. [15]. Using the experimentally measured structural fire resistance (R), the column axial capacity in fire was calculated in accordance with both DR AS/NZS 2327 and EN 1994-1-2. It was found that these code provisions can be un-conservative in some instances and that they exhibit a wide variation compared with the experimental results and modifications to the effective length recommendations as applied to continuous columns are required. 2. Experimental investigation 2.1. General The fire tests were conducted in a 2 m height × 1.5 m length × 1.5 m width furnace in IIT Roorkee, India, in accordance with EN 1364-1: 2012 [16] and with the furnace temperature controlled to match the ISO 834 [2] time-temperature curve. Fig. 2 shows the typical average furnace temperature in comparison to the ISO 834 fire curve for a typical test. The axial load was applied for 30 min before each fire test and was maintained at the applied level until the appropriate designated failure criterion from [16] was met. 2.2. Test specimens Table 1 summarises the test specimens, material properties and load
Fig. 1. Fire furnace schematic setup for fixed-pin.
Fig. 2. Measured furnace temperature (ISO 834 shown for comparison).
102
Journal of Constructional Steel Research 136 (2017) 101–109
K.U. Ukanwa et al.
Table 1 Details of test specimens. a) Plain concrete Specimen
Column size
End conditions
ƒs N/mm2
ƒc N/mm2
Eƒi,Rd kN
ηƒi
δ
λr
R Min
Failure mode
P1 P2 P5
200 mm × 200 mm × 6 mm 200 mm × 200 mm × 6 mm 220 mm × 220 mm × 6 mm
F-F F-P F-F
569 569 461
86.9 81.33 84.4
1378.4 1378.4 1415.2
0.37 0.38 0.35
0.54 0.56 0.47
0.55 0.66 0.48
37 26 38
Flexural Flexural Local
b) Conventionally reinforced concrete R1 200 mm × 200 mm × 6 mm R2 200 mm × 200 mm × 6 mm R5 220 mm × 220 mm × 6 mm
F-F F-P F-F
569 525 461
86.1 94.5 86.05
1485.6 1485.6 1604.3
0.37 0.37 0.35
0.51 0.47 0.41
0.56 0.67 0.49
46 23 72
Flexural Flexural Plastic
c) Fibre reinforced concrete F1 200 mm × 200 mm × 6 mm F2 200 mm × 200 mm × 6 mm F5 220 mm × 220 mm × 6 mm F6 220 mm × 220 mm × 6 mm
F-F F-P F-F F-P
525 525 461 461
90.9 91.9 79.86 94.3
1378.3 1378.4 1415.2 1415.2
0.37 0.37 0.36 0.33
0.51 0.51 0.49 0.45
0.54 0.66 0.47 0.59
24 25 85 51
Flexural Flexural Flexural Flexural
Note: partial safety factor = 0.9 for steel and reinforcements; 0.65 for concrete (load level only).
a) Specimen P2 prior to test
b) Load cell
c) Pin support
d) Fixed support
Fig. 3. Photograph of laboratory test setup.
1-2 [17] as follows:
levels. Two different cross-sectional dimensions of cold formed square hollow section (SHS) were tested: 200 mm × 200 mm × 6 mm and 220 mm × 220 mm × 6 mm. The square hollow sections had measured yield strengths ranging from 461 MPa to 569 MPa. The load levels shown in Table 1 are as defined in DR AS/NZS 2327 [14] and EN 1994-
ηfi, t =
Efi, d , t Rd
(1)
Fig. 3a shows a typical column in the furnace before the fire test. 103
Journal of Constructional Steel Research 136 (2017) 101–109
K.U. Ukanwa et al.
Table 2 Concrete mix proportions. Cement
Water
Coarse aggregate
Fine aggregate
Silica fumes
Super plasticizers
Steel fibre
kg/m3 500
kg/m3 150
kg/m3 1045
kg/m3 618
kg/m3 51
kg/m3 5.18
kg/m3 50
Fig. 3b shows a photograph of the hydraulic jack and load cell used. All columns had a length of 3200 mm, with fire exposure only to the middle 2000 mm. The bottom 500 mm and the top 700 mm of the column were outside the furnace and so heated only by conduction from the fire exposed length. The load was transferred using a 300 mm × 300 mm × 40 mm end plate at the top of the column. Six vent holes having a diameter of 15 mm were drilled in the column at 100 mm, 1600 mm and 3100 mm from the top of the column with three facing each other at an angle of 180 degrees to allow the build-up of steam generated during the fire to escape. From Table 1 it can be seen that six of the columns had fixed-fixed end boundary conditions while the remaining four columns had fixedpinned boundary conditions. The top ends of the column were allowed to move in the direction of the applied load, with the applied load kept constant throughout the test until failure. The pin-end boundary condition was provided through a ball (see Fig. 3c). A steel box having a 1 mm gap greater than the SHS, was welded to the steel girder to represent the fixed end support which did not allow for rotational and translational movements (see Fig. 3d).This fixed end was designed to model the fixity generated by a practical end connection for the continuous column running through to the adjacent level. There was no damage to the fixed end support after the test by visual inspection. Glass wool fibre blankets were used to seal the gap between the column and furnace to retain the heat, while allowing column movement; this was necessary to protect the sides of the furnace from damage.
a) 220 x 220 mm SHS
b) 200 x 200 mm SHS 2.3. Details of concrete infill
Fig. 4. Longitudinal reinforcement arrangement.
Three types of concrete were used for the steel tube infill, namely: plain concrete; steel fibre reinforced concrete; and conventionally reinforced concrete. Concrete placement into the SHS for each mix was conducted by filling a steel pipe with concrete and lowering the pipe into the steel tube, then withdrawing it gradually to match the rising level of concrete and, therefore, avoid segregation. A smaller pipe was used for placing the concrete into columns with the conventionally reinforced concrete core. After pouring the concrete, a poker vibrator was used to vibrate the concrete at different levels; 0.5, 1, 1.5, 2, 2.5 and 3 m from the bottom of the column to get good mixing of the concrete and removal of large bubbles of entrapped air. The filled SHS columns were stored vertically and covered at the top to allow the curing process to occur uniformly along the length of the column, by preventing moisture escape through the ends. The vent holes were temporarily sealed prior to the infill concrete being placed and this seal was removed after 5 days. Table 2 shows the mix used for the concrete infill; the coarse aggregates of metamorphic origin used were of similar geological origin to New Zealand aggregates. The specified compressive concrete cylindrical strength after 28 days was fck = 80 MPa. Table 3 shows the
concrete strength and moisture content on the test date established from 200 × 100 mm cylinders and 150 × 150 × 150 mm cubes, respectively. The concrete cube used to determine the moisture content were wrapped in a black nylon and kept in similar conditions as the test specimen. For the steel fibre reinforced concrete, Dramix hooked end steel fibre specification 5D 65/60BG were used. The length of the fibres was 60 mm, diameter was 0.9 mm and the dosage was 50 kg/m3. To improve the workability of concrete, 6 kg/m3 of polycarboxylate based super-plasticiser was also used for the steel fibre reinforced concrete mix. For the conventionally reinforced concrete infill, longitudinal reinforcement bars were tied using 6 mm diameter stirrups, shape code 51 according to BS 8666:2005 [18]. Fig. 4(a) and (b) show the arrangements of reinforcements inside the steel tubes before pouring of the concrete. The transverse reinforcement spacing was 320 mm centre to centre (c/c) and 240 mm c/c for the 220 × 220 and 200 × 200 columns, respectively.
Table 3 Concrete strength and moisture content. Column
P1
P2
P5
F1
F2
F5
F6
R1
R2
R5
ƒcm (MPa) % Moisture content Concrete age (days)
86.9 1.71 63
81.33 2.14 65
84.4 2.74 60
90.9 1.12 62
91.9 2.37 66
79.86 1.76 55
94.3 1.78 76
86.1 2.2 56
94.5 2.17 60
86.05 1.68 50
104
Journal of Constructional Steel Research 136 (2017) 101–109
K.U. Ukanwa et al.
a) Plain concrete
Fig. 5. Position of thermocouples within 200 × 200 SHS.
2.4. Instrumentation Fig. 5 shows the plan position of the thermocouples (TC) which were placed at three different levels for each column. TC1 and TC2 were positioned on the external surface of the steel tube, while TC3 to TC6 were embedded in the concrete core. TC3 was placed at 1/6 of the section width, TC4 was placed at 1/4 of the section width, TC5 was placed at 1/3 of the section width and TC6 was placed at 1/2 of the section width.
b) Conventionally reinforced concrete
2.5. Experimental results The structural fire resistance (R) of the columns are summarised in Table 1. The variation of axial deflection against time is shown in Fig. 6; the structural fire resistance is the time from commencement of test until the load bearing capacity criterion specified in EN 1363-1 was achieved (corresponding to the limiting rate of vertical contraction). Fig. 7 shows the laterally deflected shape of the columns at the ends of the tests, measured after the columns had cooled down. Three failure modes were observed, as shown in Fig. 8. A local buckling of the steel/crushing of the concrete without significant lateral deformation was observed for column P5, this occurred due to the section capacity being exceeded with concrete core squashing and steel outward local buckling before global buckling could occur. The remaining nine columns failed globally with column R5 having a plastic hinge deformation identified by a pronounced plastic hinge, while the other eight columns had a flexural buckling failure. As can be seen from Fig. 6, the initial longitudinal elongation of the steel tube in all tests is < 3 mm. This initial longitudinal elongation is much less than that previously reported in the literature, e.g [3–5,19–23]., where initial longitudinal elongation values above 15 mm are typical. The much reduced initial longitudinal elongation can be explained by the shear bond between the concrete and the steel in the unheated lengths at the top being sufficient to prevent early slip of the steel tube relative to the concrete core, while the steel tube is still sufficiently strong to take all the applied axial load on its own. From this it can be concluded that for a continuous column subjected to fire on an intermediate floor, the tendency for the column to try and expand under individual storey heating will be minimal. Columns elongating in fire due to thermal heating are subjected to an additional axial load which will cause the steel tube to lose its structural integrity earlier; this is an important consideration in design of bare steel columns [24]. As can be seen from Table 1, the steel fibre reinforced concrete
c) Fiber reinforced concrete Fig. 6. Variation of end (axial) displacement with time.
columns had a higher R when compared to the plain and bar reinforced concrete infill. From Fig. 6, it can be seen that for P5, R5 and F5, the steel fibre reinforced concrete column began its negative displacement (shortening) earlier than both the plain and conventionally reinforced concrete columns. This represents lateral deflection occurring within the heated region. The steel fibre reinforced concrete column was able to maintain its steady downward displacement and achieve a higher R due to the increased ductility and tensile strength of the concrete core, resisting the P-δ induced moments at the mid-span region of the column generated by the applied load P and the lateral deflection, δ. After the tests, the specimens were cut open to expose the concrete core. This was undertaken at mid-span or at the point of maximum curvature for the specimen which failed in plastic hinge formation (Fig. 8 (c)). Fig. 9 shows the concrete damage for specimens P5, F5 and R5. The concave side is termed the “compression” side and the convex side the “tension” side in the following description of the state of the concrete observed in each instance. As can be seen from Fig. 9(a), 105
Journal of Constructional Steel Research 136 (2017) 101–109
K.U. Ukanwa et al.
a) Fixed-fixed
b) Fixed-pinned
Fig. 7. Lateral deformation profile of the columns at end of tests.
specimen P5 exhibited crushing and spalling of the cover concrete on the “compression” side and also general crushing on the “tension” side. This indicates that the failure mode was principally crushing of the concrete core rather than the combined compression and bending failure which was exhibited by the other two infills. There were no clear transverse cracks on the tension side that are typical of tensile strain induced cracking; this despite the concrete being unreinforced. As can be seen from Fig. 9(b), specimen F5 had vertical cracks and crushing of concrete on the compression side while there was crushing of top cover concrete on the tension side. This indicates that the addition of steel fibres increased the bending moment capacity of the column at the point of failure and so changed the failure mode from compression crushing to combined compression crushing and bending. It also assisted in the concrete core withstanding the applied load for a longer period of time Compared to column P5, generating a greater R value, despite the applied load being the same in both columns. Finally, as can be seen from Fig. 9(c), specimen R5 had crushing of the cover concrete on the compression side while there was vertical and horizontal cracks with slight crushing of cover concrete on the tension side. The crack width observed in the concrete was smaller compared to those in specimen F5 due to the presence of the longitudinal reinforcement bars, which helped the concrete to withstand the applied load when the steel tube lost its load bearing capacity.
a) Local buckling (specimen P5)
3. Comparison of experimental test results with calculated axial capacity in fire
b) Overall Flexural buckling (specimen F5)
Table 4 shows, for each of the columns, the measured temperature of the steel, concrete and the reinforcement bars at the structural fire resistance (R) of the columns. The concrete core temperature shown in Table 4 is the average temperature reading from TC3 to TC6, see Fig. 5. As mentioned previously, because some of the thermocouples in the concrete were damaged, not all reinforcement bar and concrete temperature values are shown. Also shown in Table 4 are the temperatures calculated in accordance with DR AS/NZS 2327 [14]. As can be seen in Table 4, for the steel temperatures, the equivalent temperature predicted by DR AS/NZS 2327 has a maximum difference of about 4% to the measured temperatures for all concrete infill types. On the other hand, for the concrete core and longitudinal reinforcement bars, the temperatures calculated in accordance with DR AS/NZS 2327 are conservative by, on average, 12% for standard fire tests. Using the temperatures calculated in accordance with DR AS/NZS
c) Plastic hinge formation (specimen R5) Fig. 8. Photograph of columns at failure.
106
Journal of Constructional Steel Research 136 (2017) 101–109
K.U. Ukanwa et al.
Fig. 9. Photograph showing concrete damage at failure region.
2327, Table 5 compares the experimental fire axial loads against those calculated in accordance with DR AS/NZS 2327 and advanced calculation method given in EN 1994-1-2. It should be noted that the effective length factors are different in DR AS/NZS 2327 and EN 1994-1-2. For the use of DR AS/NZS 2327, the values of k are 0.7 and 0.85 for fixedfixed and fixed-pinned, respectively, reflecting the Australasian practice to use actual rather than ideal values of column end rotational fixity On the other hand, for the Eurocodes, including EN 1994-1-2, the ideal rotational end fixity is used for both second order effect determination and for member. As can be seen from Table 5(a), for the plain concrete specimens, the axial capacity in fire predicted by EN 1994-1-2 are un-conservative Compared with the experimental test results. For specimen P5 the value of Efi,Exp/Efi,Rd. is 0.52 while for specimen P1 the value of Efi,Exp/Efi,Rd. is 0.66. On the other hand, while the values of Efi,Exp/Efi,Rd. for DR AS/ NZS 2327 are all higher, for specimen P5 the value of Efi,Exp/Efi,Rd. is 0.74 and so remains un-conservative. While the difference in the values of Efi,Exp/Efi,Rd. between the two codes can be attributed to their different effective length factors k, as mentioned previously, the deformed geometry between the fixed-fixed and fixed-pinned were found to be similar, with the exception of column R5 that had a plastic failure mode (see Fig. 7). As can be seen from Table 5(b), the same observations can be made for the conventionally reinforced concrete specimens. The value of Efi,Exp/Efi,Rd. for DR AS/NZS 2327 are all lower than those for EN 19941-2. For DR AS/NZS 2327, only specimen R2 has a value of Efi,Exp/Efi,Rd. that is un-conservative. It should be noted that only for specimen R2, the value of R is < 30 min. Table 5(c) shows the same results for the fibre reinforced concrete specimens. For DR AS/NZS 2327, specimens F1 and F2 can be seen to be conservative and again both the specimens had a value of R < 30 min.
a) Specimen P5
3.1. Sensitivity to value of K of 0.9. The values of Efi,Exp/Efi,Rd. can be seen to be sensitive to the value of the effective length factor “k” and also to the value of R. While there is insufficient test data for an overall conclusion to be drawn, it is of interest to compare the results again if a more conservative value of “k” of 0.9 is adopted for both fixed-fixed and fixed-pinned. This value is obtained from looking at the lateral deformed shape of the tested specimens (see Fig. 7) and applying the first principles of the effective length being “the length of an equivalent pin ended column”. Table 6 shows the same results calculated in accordance with DR AS/NZS 2327 for all three types of infill, with the exception of conventionally reinforced and fibre reinforced concrete infill where R > 30 min. From Table 6, the value of Efi,Exp/Efi,Rd. for all specimens filled with plain concrete are now conservative. For specimens filled with steel fibre reinforced concrete and conventionally reinforced concrete, the value of Efi,Exp/Efi,Rd. are only slightly un-conservative and so are satisfactory to an acceptable degree of tolerance. For comparison, Table 6 also shows the same values of Efi,Exp/Efi,Rd. calculated using the above recommendations for the effective length factor but calculated in accordance with EN 1994-1-2. The value of Efi,Exp/Efi,Rd. calculated in accordance with EN 1994-1-2 are only around 4% higher than those calculated in accordance with DR AS/NZS 2327.
b) Specimen F5
4. Conclusions This section presents the overall conclusions that can be made for continuous columns where the floors above and below are cooler. These conclusions are drawn from ten experimental tests, with different concrete infills (namely, plain concrete, bar reinforced concrete, and steel fibre reinforced concrete). It should be noted that these infills have
c) Specimen R5 107
Journal of Constructional Steel Research 136 (2017) 101–109
K.U. Ukanwa et al.
Table 4 Comparison of experimentally recorded temperatures with values calculated in accordance with DR AS/NZS 2327 a) Plain concrete Specimen
P1 P2 P5
R
Am/V
Experimental
DR AS/NZS 2327
mins
m− 1
o
Ts C
Tc o C
Ts o C
Tc o C
37 26 38
20 20 18
664 658 774
472 242 408
740 658 742
418 342 397
b) Conventionally reinforced concrete Specimen
R1 R2 R5
R
Am/V
Experimental
DR AS/NZS 2327
mins
m− 1
Ts o C
Tc o C
Tr o C
Ts o C
Tc o C
Tr o C
46 23 72
20 20 18
770 676 929
– 172 –
– 156 –
800 633 928
476 320 581
376 185 529
c) Fibre reinforced concrete Specimen
F1 F2 F5 F6
R
Am/V
Experimental
DR AS/NZS 2327
mins
m− 1
Ts o C
Tc o C
Ts o C
Tc o C
24 25 85 51
20 20 18 18
666 647 963 693
292 214 – 395
641 650 972 825
328 335 637 474
Note: “–” means un-reliable data. Note: “-” means un-reliable data. Table 5 Comparison of experimental fire axial loads with values calculated from DR AS/NZS 2327 and EN 1994-1-2. a) Plain concrete Specimen
DR AS/NZS 2327
R
Eƒi,Exp
mins
kN
37 26 38
1378 1378 1415
0.7 0.85 0.7
b) Conventionally reinforced concrete R1 46 1485 R2 23 1485 R5 72 1604 c) Fibre reinforced concrete F1 24 F2 25 F5 85 F6 51
Eƒi,Rd kN
Eƒi,Exp
1.29 1.44 1.10
1311 1312 1917
1.05 1.05 0.74
0.5 0.7 0.5
0.7 0.85 0.7
1.36 1.42 1.39
1180 1723 1042
1.26 0.86 1.53
0.7 0.85 0.7 0.85
1.18 1.43 1.14 1.14
1968 1478 760 1167
0.5 0.93 1.86 1.21
k
P1 P2 P5
1378 1378 1415 1415
λT
EN 1994-1-2 k
λT
Eƒi,Rd kN
Eƒi,Exp
0.94 1.21 0.80
2068 1809 2699
0.66 0.76 0.52
0.5 0.7 0.5
0.98 1.19 1.01
1927 2380 1892
0.77 0.62 0.84
0.5 0.7 0.5 0.5
0.83 1.19 0.98 1.29
2895 2040 1095 1630
0.47 0.67 1.29 0.87
Eƒi,Rd
Eƒi,Rd
0.49 are conservative when the “k” values as specified in DR AS/NZS 2327 are applied. Therefore, the following conclusion can be drawn:
a measured concrete compressive strength ranging from 80 MPa to 95 MPa. All tests were conducted on square hollow steel sections having nominal cross-sections dimensions of 200 mm × 200 mm × 6 mm and 220 mm × 220 mm × 6 mm; the yield strengths of the sections tested ranged from 461 MPa to 569 MPa. In the tests, end restraint conditions of fixed-fixed (F-F) and pinned-fixed (P-F)) with considered. It was found that CFST columns filled with conventional reinforced concrete having a fixed-fixed boundary condition and columns filled with steel fibre reinforced concrete having steel contribution ratio of 0.47 and
1. A comparison of the experimental axial capacity in fire with that determined by calculation to DR AS/NZS 2327 and EN 1994-1-2 has shown that the standards can be un-conservative, particularly for columns filled with plain concrete and with an R value of < 30 min. For these cases, an effective length factor value of 0.9 has been shown to be more representative for continuous CFST column where
108
Journal of Constructional Steel Research 136 (2017) 101–109
K.U. Ukanwa et al.
References
Table 6 Comparison of experimental fire axial load using “k” factor of 0.9.
[1] M. Yu, X. Zha, J. Ye, B. Wang, A unified method for calculating fire resistance of solid and hollow concrete-filled steel tube columns based on average temperature, Eng.Struct. 71 (2014) 12–22. [2] ISO I. 834, Fire Resistance Tests-elements of Building Construction, International Organization for Standardization, Geneva, Switzerland, 1999. [3] M.L. Romero, V. Moliner, A. Espinos, C. Ibañez, A. Hospitaler, Fire behavior of axially loaded slender high strength concrete-filled tubular columns, J. Constr. Steel Res. 67 (12) (2011) 1953–1965. [4] V.K.R. Kodur, J. Latour, Experimental Studies on the Fire Resistance of Hollow Steel Columns Filled With High-strength Concrete, (2005). [5] T. Lie, R. Irwin, Fire resistance of rectangular steel columns filled with barreinforced concrete, J.Struct.Eng 121 (5) (1995) 797–805. [6] N. Mago, S.J. Hicks, Fire behaviour of slender, highly utilized, eccentrically loaded concrete filled tubular columns, J. Constr. Steel Res. 119 (2016) 123–132. [7] A. Espinos, M.L. Romero, A. Hospitaler, Advanced model for predicting the fire response of concrete filled tubular columns, J. Constr. Steel Res. 66 (8) (2010) 1030–1046. [8] Z. Wang, J. Wu, J. Wang, Experimental and numerical analysis on effect of fibre aspect ratio on mechanical properties of SRFC, Constr. Build. Mater. 24 (4) (2010) 559–565. [9] P. Schaumann, V. Kodur, O. Bahr, Fire behaviour of hollow structural section steel columns filled with high strength concrete, J. Constr. Steel Res. 65 (8) (2009) 1794–1802. [10] V. Moliner, A. Espinos, M.L. Romero, A. Hospitaler, Fire behavior of eccentrically loaded slender high strength concrete-filled tubular columns, J. Constr. Steel Res. 83 (2013) 137–146. [11] V.K. Kodur, R. Fike, Response of concrete-filled HSS columns in real fires, Eng. J. AISC Inc. 46 (4) (2009) 243–256. [12] A. Espinos, M.L. Romero, E. Serra, A. Hospitaler, Experimental investigation on the fire behaviour of rectangular and elliptical slender concrete-filled tubular columns, Thin-Walled Struct. 93 (2015) 137–148. [13] A. Espinos, M.L. Romero, E. Serra, A. Hospitaler, Circular and square slender concrete-filled tubular columns under large eccentricities and fire, J. Constr. Steel Res. 110 (2015) 90–100. [14] DR AS/NZS 2327, Composite Steel-concrete Construction for Buildings, Australian/ New Zealand Standards, 2016. [15] A. Espinos, M.L. Romero, A. Hospitaler, Simple calculation model for evaluating the fire resistance of unreinforced concrete filled tubular columns, Eng. Struct. 42 (2012) 231–244. [16] EN B. 1363–1, Fire Resistance Tests. General Requirements. BSI Standard 2012, (2012). [17] Eurocode E, 4: Design of composite steel and concrete structures–part 1–2: general rules–structural fire design, British Standards Institution, BS EN, 1994, pp. 1–2. [18] British Standards Institution, BS 8666: 2005: scheduling, dimensioning, bending and cutting of steel reinforcement for concrete, Specification: British Standards Institution, 2005. [19] M. Chabot, T.T. Lie, Experimental Studies on the Fire Resistance of Hollow Steel Columns Filled with Bar-reinforced Concrete, (1992). [20] J. Ding, Y. Wang, Realistic modelling of thermal and structural behaviour of unprotected concrete filled tubular columns in fire, J. Constr. Steel Res. 64 (10) (2008) 1086–1102. [21] S. Hong, A.H. Varma, Analytical modeling of the standard fire behavior of loaded CFT columns, J. Constr. Steel Res. 65 (1) (2009) 54–69. [22] K. Chung, S. Park, S. Choi, Fire resistance of concrete filled square steel tube columns subjected to eccentric axial load, Int. J. Steel Struct. 9 (1) (2009) 69–76. [23] J. Yin, X. Zha, L. Li, Fire resistance of axially loaded concrete filled steel tube columns, J. Constr. Steel Res. 62 (7) (2006) 723–729. [24] M. Spearpoint (Ed.), Fire Engineering Design Guide, third ed., Centre for Advanced Engineering: Christchurch, New Zealand, 2008.
a) Plain concrete Specimen
R
Eƒi,Exp
mins
kN
37 26 38
1378 1378 1415
DR AS/NZS 2327 λT
EN 1994-1-2
Eƒi,Rd kN
Eƒi,Exp
889 1199 1370
1.55 1.15 1.07
b) Conventionally reinforced concrete R1 46 1485 1.36 1180 R2 23 1485 1.49 1577 R5 72 1604 1.39 1042 c) Fibre reinforced concrete F1 24 1378 F2 25 1378 F5 85 1415 F6 51 1415
P1 P2 P5
1.63 1.51 1.39
1.48 1.50 1.14 1.44
1389 1351 760 1166
λT
Eƒi,Rd kN
Eƒi,Exp
1.63 1.51 1.39
919 1243 1373
1.50 1.11 1.03
1.26 0.94 1.54
1.36 1.49 1.39
1232 1636 1193
1.2 0.91 1.34
0.99 1.02 1.86 1.21
1.48 1.50 1.14 1.14
1440 1402 806 1208
0.96 0.98 1.75 1.17
Eƒi,Rd
Eƒi,Rd
the member temperatures are calculated in accordance to DR AS/ NZS 2327; based on the experimentally observed effective length of these columns. This effective length factor was suggested from examination of the lateral deformed shape of the tested specimens. 2. For columns having slenderness ranging from 0.47 to 0.67 in the ambient temperature condition, and with percentage of bar reinforcement in accordance to EN 1994-1-2, it was found that columns filled with steel fibre reinforced concrete had the highest structural fire resistance to the concentric compression loaded columns. This was due to the even distribution of steel fibre in the concrete core, which increased the tensile strength of the concrete core and so increased the bending resistance. 3. The longitudinal elongation of the steel was < 3 mm. This is much less than the longitudinal elongation recorded in other CFST fire tests where the entire height was heated. This was as a result of the shear bond between the steel and concrete at the cooler top and bottom region of the steel which restrained the heated part from further elongation.
Acknowledgements The authors wish to express gratitude to the New Zealand Heavy Engineering Educational & Research Foundation (HEERF) for their scholarship support, also to the department of Civil Engineering, IIT Roorkee for providing the laboratory used for the experiments.
109
Contents lists available at ScienceDirect
Journal of Constructional Steel Research journal homepage: www.elsevier.com/locate/jcsr
Behaviour of continuous concrete filled steel tubular columns loaded concentrically in fire
MARK
Kingsley U. Ukanwaa, Umesh Sharmab, Stephen J. Hicksc, Anthony Abud, James B.P. Lima, G. Charles Cliftona,⁎ a
Department of Civil and Environmental Engineering, The University of Auckland, Auckland, New Zealand Department of Civil Engineering, India Institute of Technology, Roorkee, India c New Zealand Heavy Engineering Research Association, HERA House, Auckland, New Zealand d Department of Civil and Natural Resources Engineering, University of Canterbury, Christchurch, New Zealand b
A R T I C L E I N F O
A B S T R A C T
Keywords: Fire resistance Concrete-filled steel tubular columns Continuous columns Fibre reinforced concrete
Design recommendations for concrete filled steel tubular (CFST) columns in fire are based on the results of experimental standard fire testing of CFST members where the same temperature is applied to the column over the full column height. However, this is not representative of a CFST column in a typical building, which is continuous between floors and which, in fire, is subjected to severe fire conditions on one floor at a time while the floors above and below remain cooler. In the experimental tests described in this paper, the columns are of 3.2 m height with the fire applied only to the central 2 m. Significant differences are observed between these tests and those previously conducted due to the partial length heating. In total, ten tests are conducted; the tests cover three different types of infill: plain concrete; bar reinforced concrete; and steel fibre reinforced concrete. End restraint conditions of fixed-fixed (F-F) and pinned-fixed (P-F) are considered; the axial load levels are between 0.33 to 0.38 of the squash load. The longitudinal elongation of the steel tube was less than 3 mm. Using the experimentally measured structural fire resistance (R), the axial capacity in fire was calculated in accordance with the codes of practice and are Compared with the experimentally tested structural fire resistance, showing that in some instances current design practice can be un-conservative.
1. Introduction Unprotected concrete filled steel tubes (CFST) (i.e. with no insulation material applied to the outside of the steel tube) can have an inherent fire resistance, due to concrete preventing inward buckling of the steel tube and hence increasing its local and global buckling resistance and the steel tube partially or totally confining the concrete, to give the column a higher compressive strength. In addition, the steel tube acts as a partial radiation shield to the concrete [1]. In the literature, the performance of CFST under ISO 834 [2] Standard Fire conditions has been considered by, amongst others, Romero et al. [3], Kodur and Latour [4] and Lie and Irwin [5], and has covered three types of in-fill; plain concrete, steel fibre reinforced concrete and rebar reinforced concrete. Finite element analyses studies have also been reported [6–9] which cover CFST columns filled with normal strength and high strength concrete loaded axially with high utilization factor in fire. The finite element analysis studies, however, have not covered continuous columns in fire.
⁎
Corresponding author. E-mail address: [email protected] (G.C. Clifton).
http://dx.doi.org/10.1016/j.jcsr.2017.05.011 Received 16 January 2017; Received in revised form 10 May 2017; Accepted 12 May 2017 0143-974X/ © 2017 Elsevier Ltd. All rights reserved.
In all previously reported experimental tests and finite element analyses, the entire column height is heated. In New Zealand, these columns are required to be continuous over their full length for seismic design requirements. The full length heating of the column is not representative of a fire cell in a continuous column construction where the fire is on one floor and the floors above and below remain cooler. This paper considers the behaviour of a continuous column heated over part of its length at one time. The experimental tests reported herein consider the issue of partially heated length in a continuous column and different concrete infills having measured compressive strength ranging from 80 MPa to 95 MPa. In total ten tests were conducted, with temperature only applied to 2 m of a 3.2 m high column (see Fig. 1); all tests were conducted on square hollow steel sections having nominal crosssections dimensions of 200 mm × 200 mm × 6 mm and 220 mm × 220 mm × 6 mm; the yield strengths of the sections tested ranged from 461 MPa to 569 MPa. The tests cover three different types of infill: plain concrete; bar reinforced concrete; and steel fibre
Journal of Constructional Steel Research 136 (2017) 101–109
K.U. Ukanwa et al.
Nomenclature As,T Ac,T Ar,T Am/V Es,T Ec,T Eƒi,Exp Eƒi,Rd Ec,sec,T Er,T (EI)fi F-F F-P ƒc ƒr ƒs
Ic,T Is,T
cross-sectional area of steel profile at the temperature °c cross-sectional area of concrete at the temperature °c cross-sectional area of reinforcement at the temperature °c section factor modulus of elasticity of steel tangent modulus of concrete at the temperature °c design effect of actions in fire situation for laboratory experiment design effect of actions in fire situation for design code secant modulus of concrete at the temperature °c modulus of elasticity of reinforcement at the temperature °c effective flexural stiffness in fire situation Fixed - Fixed Fixed - Pinned compressive cylinder strength of concrete at room temperature yield strength of reinforcement at room temperature yield strength of structural steel at room temperature
Ir,T k Le,T Nc,ƒi,Rd Nb,ƒi,Rd.,t R t αc T Tc Ts Tr λr λ,T ηƒi δ
reinforced concrete [10–13]. End restraint conditions of fixed-fixed (FF) and pinned-fixed (P-F) were considered. Compressive load levels ηfi of 0.33 and 0.38 were applied, with the loads determined in accordance
second moment of area of concrete at the temperature °c second moment of area of steel profile at the temperature °c second moment of area of reinforcement at the temperature °c effective length factor buckling length of column in fire situation design cross-sectional plastic resistance to axial compression in fire design axial buckling load of column in fire situation structural fire resistance steel tube thickness member slenderness reduction factor temperature temperature of concrete temperature of steel temperature of reinforcement relative slenderness of column at room temperature relative slenderness of column in fire situation design load level in fire condition steel contribution ratio
with the provisions in DR AS/NZS 2327 [14], which are based on the recommendations of Espinos et al. [15]. Using the experimentally measured structural fire resistance (R), the column axial capacity in fire was calculated in accordance with both DR AS/NZS 2327 and EN 1994-1-2. It was found that these code provisions can be un-conservative in some instances and that they exhibit a wide variation compared with the experimental results and modifications to the effective length recommendations as applied to continuous columns are required. 2. Experimental investigation 2.1. General The fire tests were conducted in a 2 m height × 1.5 m length × 1.5 m width furnace in IIT Roorkee, India, in accordance with EN 1364-1: 2012 [16] and with the furnace temperature controlled to match the ISO 834 [2] time-temperature curve. Fig. 2 shows the typical average furnace temperature in comparison to the ISO 834 fire curve for a typical test. The axial load was applied for 30 min before each fire test and was maintained at the applied level until the appropriate designated failure criterion from [16] was met. 2.2. Test specimens Table 1 summarises the test specimens, material properties and load
Fig. 1. Fire furnace schematic setup for fixed-pin.
Fig. 2. Measured furnace temperature (ISO 834 shown for comparison).
102
Journal of Constructional Steel Research 136 (2017) 101–109
K.U. Ukanwa et al.
Table 1 Details of test specimens. a) Plain concrete Specimen
Column size
End conditions
ƒs N/mm2
ƒc N/mm2
Eƒi,Rd kN
ηƒi
δ
λr
R Min
Failure mode
P1 P2 P5
200 mm × 200 mm × 6 mm 200 mm × 200 mm × 6 mm 220 mm × 220 mm × 6 mm
F-F F-P F-F
569 569 461
86.9 81.33 84.4
1378.4 1378.4 1415.2
0.37 0.38 0.35
0.54 0.56 0.47
0.55 0.66 0.48
37 26 38
Flexural Flexural Local
b) Conventionally reinforced concrete R1 200 mm × 200 mm × 6 mm R2 200 mm × 200 mm × 6 mm R5 220 mm × 220 mm × 6 mm
F-F F-P F-F
569 525 461
86.1 94.5 86.05
1485.6 1485.6 1604.3
0.37 0.37 0.35
0.51 0.47 0.41
0.56 0.67 0.49
46 23 72
Flexural Flexural Plastic
c) Fibre reinforced concrete F1 200 mm × 200 mm × 6 mm F2 200 mm × 200 mm × 6 mm F5 220 mm × 220 mm × 6 mm F6 220 mm × 220 mm × 6 mm
F-F F-P F-F F-P
525 525 461 461
90.9 91.9 79.86 94.3
1378.3 1378.4 1415.2 1415.2
0.37 0.37 0.36 0.33
0.51 0.51 0.49 0.45
0.54 0.66 0.47 0.59
24 25 85 51
Flexural Flexural Flexural Flexural
Note: partial safety factor = 0.9 for steel and reinforcements; 0.65 for concrete (load level only).
a) Specimen P2 prior to test
b) Load cell
c) Pin support
d) Fixed support
Fig. 3. Photograph of laboratory test setup.
1-2 [17] as follows:
levels. Two different cross-sectional dimensions of cold formed square hollow section (SHS) were tested: 200 mm × 200 mm × 6 mm and 220 mm × 220 mm × 6 mm. The square hollow sections had measured yield strengths ranging from 461 MPa to 569 MPa. The load levels shown in Table 1 are as defined in DR AS/NZS 2327 [14] and EN 1994-
ηfi, t =
Efi, d , t Rd
(1)
Fig. 3a shows a typical column in the furnace before the fire test. 103
Journal of Constructional Steel Research 136 (2017) 101–109
K.U. Ukanwa et al.
Table 2 Concrete mix proportions. Cement
Water
Coarse aggregate
Fine aggregate
Silica fumes
Super plasticizers
Steel fibre
kg/m3 500
kg/m3 150
kg/m3 1045
kg/m3 618
kg/m3 51
kg/m3 5.18
kg/m3 50
Fig. 3b shows a photograph of the hydraulic jack and load cell used. All columns had a length of 3200 mm, with fire exposure only to the middle 2000 mm. The bottom 500 mm and the top 700 mm of the column were outside the furnace and so heated only by conduction from the fire exposed length. The load was transferred using a 300 mm × 300 mm × 40 mm end plate at the top of the column. Six vent holes having a diameter of 15 mm were drilled in the column at 100 mm, 1600 mm and 3100 mm from the top of the column with three facing each other at an angle of 180 degrees to allow the build-up of steam generated during the fire to escape. From Table 1 it can be seen that six of the columns had fixed-fixed end boundary conditions while the remaining four columns had fixedpinned boundary conditions. The top ends of the column were allowed to move in the direction of the applied load, with the applied load kept constant throughout the test until failure. The pin-end boundary condition was provided through a ball (see Fig. 3c). A steel box having a 1 mm gap greater than the SHS, was welded to the steel girder to represent the fixed end support which did not allow for rotational and translational movements (see Fig. 3d).This fixed end was designed to model the fixity generated by a practical end connection for the continuous column running through to the adjacent level. There was no damage to the fixed end support after the test by visual inspection. Glass wool fibre blankets were used to seal the gap between the column and furnace to retain the heat, while allowing column movement; this was necessary to protect the sides of the furnace from damage.
a) 220 x 220 mm SHS
b) 200 x 200 mm SHS 2.3. Details of concrete infill
Fig. 4. Longitudinal reinforcement arrangement.
Three types of concrete were used for the steel tube infill, namely: plain concrete; steel fibre reinforced concrete; and conventionally reinforced concrete. Concrete placement into the SHS for each mix was conducted by filling a steel pipe with concrete and lowering the pipe into the steel tube, then withdrawing it gradually to match the rising level of concrete and, therefore, avoid segregation. A smaller pipe was used for placing the concrete into columns with the conventionally reinforced concrete core. After pouring the concrete, a poker vibrator was used to vibrate the concrete at different levels; 0.5, 1, 1.5, 2, 2.5 and 3 m from the bottom of the column to get good mixing of the concrete and removal of large bubbles of entrapped air. The filled SHS columns were stored vertically and covered at the top to allow the curing process to occur uniformly along the length of the column, by preventing moisture escape through the ends. The vent holes were temporarily sealed prior to the infill concrete being placed and this seal was removed after 5 days. Table 2 shows the mix used for the concrete infill; the coarse aggregates of metamorphic origin used were of similar geological origin to New Zealand aggregates. The specified compressive concrete cylindrical strength after 28 days was fck = 80 MPa. Table 3 shows the
concrete strength and moisture content on the test date established from 200 × 100 mm cylinders and 150 × 150 × 150 mm cubes, respectively. The concrete cube used to determine the moisture content were wrapped in a black nylon and kept in similar conditions as the test specimen. For the steel fibre reinforced concrete, Dramix hooked end steel fibre specification 5D 65/60BG were used. The length of the fibres was 60 mm, diameter was 0.9 mm and the dosage was 50 kg/m3. To improve the workability of concrete, 6 kg/m3 of polycarboxylate based super-plasticiser was also used for the steel fibre reinforced concrete mix. For the conventionally reinforced concrete infill, longitudinal reinforcement bars were tied using 6 mm diameter stirrups, shape code 51 according to BS 8666:2005 [18]. Fig. 4(a) and (b) show the arrangements of reinforcements inside the steel tubes before pouring of the concrete. The transverse reinforcement spacing was 320 mm centre to centre (c/c) and 240 mm c/c for the 220 × 220 and 200 × 200 columns, respectively.
Table 3 Concrete strength and moisture content. Column
P1
P2
P5
F1
F2
F5
F6
R1
R2
R5
ƒcm (MPa) % Moisture content Concrete age (days)
86.9 1.71 63
81.33 2.14 65
84.4 2.74 60
90.9 1.12 62
91.9 2.37 66
79.86 1.76 55
94.3 1.78 76
86.1 2.2 56
94.5 2.17 60
86.05 1.68 50
104
Journal of Constructional Steel Research 136 (2017) 101–109
K.U. Ukanwa et al.
a) Plain concrete
Fig. 5. Position of thermocouples within 200 × 200 SHS.
2.4. Instrumentation Fig. 5 shows the plan position of the thermocouples (TC) which were placed at three different levels for each column. TC1 and TC2 were positioned on the external surface of the steel tube, while TC3 to TC6 were embedded in the concrete core. TC3 was placed at 1/6 of the section width, TC4 was placed at 1/4 of the section width, TC5 was placed at 1/3 of the section width and TC6 was placed at 1/2 of the section width.
b) Conventionally reinforced concrete
2.5. Experimental results The structural fire resistance (R) of the columns are summarised in Table 1. The variation of axial deflection against time is shown in Fig. 6; the structural fire resistance is the time from commencement of test until the load bearing capacity criterion specified in EN 1363-1 was achieved (corresponding to the limiting rate of vertical contraction). Fig. 7 shows the laterally deflected shape of the columns at the ends of the tests, measured after the columns had cooled down. Three failure modes were observed, as shown in Fig. 8. A local buckling of the steel/crushing of the concrete without significant lateral deformation was observed for column P5, this occurred due to the section capacity being exceeded with concrete core squashing and steel outward local buckling before global buckling could occur. The remaining nine columns failed globally with column R5 having a plastic hinge deformation identified by a pronounced plastic hinge, while the other eight columns had a flexural buckling failure. As can be seen from Fig. 6, the initial longitudinal elongation of the steel tube in all tests is < 3 mm. This initial longitudinal elongation is much less than that previously reported in the literature, e.g [3–5,19–23]., where initial longitudinal elongation values above 15 mm are typical. The much reduced initial longitudinal elongation can be explained by the shear bond between the concrete and the steel in the unheated lengths at the top being sufficient to prevent early slip of the steel tube relative to the concrete core, while the steel tube is still sufficiently strong to take all the applied axial load on its own. From this it can be concluded that for a continuous column subjected to fire on an intermediate floor, the tendency for the column to try and expand under individual storey heating will be minimal. Columns elongating in fire due to thermal heating are subjected to an additional axial load which will cause the steel tube to lose its structural integrity earlier; this is an important consideration in design of bare steel columns [24]. As can be seen from Table 1, the steel fibre reinforced concrete
c) Fiber reinforced concrete Fig. 6. Variation of end (axial) displacement with time.
columns had a higher R when compared to the plain and bar reinforced concrete infill. From Fig. 6, it can be seen that for P5, R5 and F5, the steel fibre reinforced concrete column began its negative displacement (shortening) earlier than both the plain and conventionally reinforced concrete columns. This represents lateral deflection occurring within the heated region. The steel fibre reinforced concrete column was able to maintain its steady downward displacement and achieve a higher R due to the increased ductility and tensile strength of the concrete core, resisting the P-δ induced moments at the mid-span region of the column generated by the applied load P and the lateral deflection, δ. After the tests, the specimens were cut open to expose the concrete core. This was undertaken at mid-span or at the point of maximum curvature for the specimen which failed in plastic hinge formation (Fig. 8 (c)). Fig. 9 shows the concrete damage for specimens P5, F5 and R5. The concave side is termed the “compression” side and the convex side the “tension” side in the following description of the state of the concrete observed in each instance. As can be seen from Fig. 9(a), 105
Journal of Constructional Steel Research 136 (2017) 101–109
K.U. Ukanwa et al.
a) Fixed-fixed
b) Fixed-pinned
Fig. 7. Lateral deformation profile of the columns at end of tests.
specimen P5 exhibited crushing and spalling of the cover concrete on the “compression” side and also general crushing on the “tension” side. This indicates that the failure mode was principally crushing of the concrete core rather than the combined compression and bending failure which was exhibited by the other two infills. There were no clear transverse cracks on the tension side that are typical of tensile strain induced cracking; this despite the concrete being unreinforced. As can be seen from Fig. 9(b), specimen F5 had vertical cracks and crushing of concrete on the compression side while there was crushing of top cover concrete on the tension side. This indicates that the addition of steel fibres increased the bending moment capacity of the column at the point of failure and so changed the failure mode from compression crushing to combined compression crushing and bending. It also assisted in the concrete core withstanding the applied load for a longer period of time Compared to column P5, generating a greater R value, despite the applied load being the same in both columns. Finally, as can be seen from Fig. 9(c), specimen R5 had crushing of the cover concrete on the compression side while there was vertical and horizontal cracks with slight crushing of cover concrete on the tension side. The crack width observed in the concrete was smaller compared to those in specimen F5 due to the presence of the longitudinal reinforcement bars, which helped the concrete to withstand the applied load when the steel tube lost its load bearing capacity.
a) Local buckling (specimen P5)
3. Comparison of experimental test results with calculated axial capacity in fire
b) Overall Flexural buckling (specimen F5)
Table 4 shows, for each of the columns, the measured temperature of the steel, concrete and the reinforcement bars at the structural fire resistance (R) of the columns. The concrete core temperature shown in Table 4 is the average temperature reading from TC3 to TC6, see Fig. 5. As mentioned previously, because some of the thermocouples in the concrete were damaged, not all reinforcement bar and concrete temperature values are shown. Also shown in Table 4 are the temperatures calculated in accordance with DR AS/NZS 2327 [14]. As can be seen in Table 4, for the steel temperatures, the equivalent temperature predicted by DR AS/NZS 2327 has a maximum difference of about 4% to the measured temperatures for all concrete infill types. On the other hand, for the concrete core and longitudinal reinforcement bars, the temperatures calculated in accordance with DR AS/NZS 2327 are conservative by, on average, 12% for standard fire tests. Using the temperatures calculated in accordance with DR AS/NZS
c) Plastic hinge formation (specimen R5) Fig. 8. Photograph of columns at failure.
106
Journal of Constructional Steel Research 136 (2017) 101–109
K.U. Ukanwa et al.
Fig. 9. Photograph showing concrete damage at failure region.
2327, Table 5 compares the experimental fire axial loads against those calculated in accordance with DR AS/NZS 2327 and advanced calculation method given in EN 1994-1-2. It should be noted that the effective length factors are different in DR AS/NZS 2327 and EN 1994-1-2. For the use of DR AS/NZS 2327, the values of k are 0.7 and 0.85 for fixedfixed and fixed-pinned, respectively, reflecting the Australasian practice to use actual rather than ideal values of column end rotational fixity On the other hand, for the Eurocodes, including EN 1994-1-2, the ideal rotational end fixity is used for both second order effect determination and for member. As can be seen from Table 5(a), for the plain concrete specimens, the axial capacity in fire predicted by EN 1994-1-2 are un-conservative Compared with the experimental test results. For specimen P5 the value of Efi,Exp/Efi,Rd. is 0.52 while for specimen P1 the value of Efi,Exp/Efi,Rd. is 0.66. On the other hand, while the values of Efi,Exp/Efi,Rd. for DR AS/ NZS 2327 are all higher, for specimen P5 the value of Efi,Exp/Efi,Rd. is 0.74 and so remains un-conservative. While the difference in the values of Efi,Exp/Efi,Rd. between the two codes can be attributed to their different effective length factors k, as mentioned previously, the deformed geometry between the fixed-fixed and fixed-pinned were found to be similar, with the exception of column R5 that had a plastic failure mode (see Fig. 7). As can be seen from Table 5(b), the same observations can be made for the conventionally reinforced concrete specimens. The value of Efi,Exp/Efi,Rd. for DR AS/NZS 2327 are all lower than those for EN 19941-2. For DR AS/NZS 2327, only specimen R2 has a value of Efi,Exp/Efi,Rd. that is un-conservative. It should be noted that only for specimen R2, the value of R is < 30 min. Table 5(c) shows the same results for the fibre reinforced concrete specimens. For DR AS/NZS 2327, specimens F1 and F2 can be seen to be conservative and again both the specimens had a value of R < 30 min.
a) Specimen P5
3.1. Sensitivity to value of K of 0.9. The values of Efi,Exp/Efi,Rd. can be seen to be sensitive to the value of the effective length factor “k” and also to the value of R. While there is insufficient test data for an overall conclusion to be drawn, it is of interest to compare the results again if a more conservative value of “k” of 0.9 is adopted for both fixed-fixed and fixed-pinned. This value is obtained from looking at the lateral deformed shape of the tested specimens (see Fig. 7) and applying the first principles of the effective length being “the length of an equivalent pin ended column”. Table 6 shows the same results calculated in accordance with DR AS/NZS 2327 for all three types of infill, with the exception of conventionally reinforced and fibre reinforced concrete infill where R > 30 min. From Table 6, the value of Efi,Exp/Efi,Rd. for all specimens filled with plain concrete are now conservative. For specimens filled with steel fibre reinforced concrete and conventionally reinforced concrete, the value of Efi,Exp/Efi,Rd. are only slightly un-conservative and so are satisfactory to an acceptable degree of tolerance. For comparison, Table 6 also shows the same values of Efi,Exp/Efi,Rd. calculated using the above recommendations for the effective length factor but calculated in accordance with EN 1994-1-2. The value of Efi,Exp/Efi,Rd. calculated in accordance with EN 1994-1-2 are only around 4% higher than those calculated in accordance with DR AS/NZS 2327.
b) Specimen F5
4. Conclusions This section presents the overall conclusions that can be made for continuous columns where the floors above and below are cooler. These conclusions are drawn from ten experimental tests, with different concrete infills (namely, plain concrete, bar reinforced concrete, and steel fibre reinforced concrete). It should be noted that these infills have
c) Specimen R5 107
Journal of Constructional Steel Research 136 (2017) 101–109
K.U. Ukanwa et al.
Table 4 Comparison of experimentally recorded temperatures with values calculated in accordance with DR AS/NZS 2327 a) Plain concrete Specimen
P1 P2 P5
R
Am/V
Experimental
DR AS/NZS 2327
mins
m− 1
o
Ts C
Tc o C
Ts o C
Tc o C
37 26 38
20 20 18
664 658 774
472 242 408
740 658 742
418 342 397
b) Conventionally reinforced concrete Specimen
R1 R2 R5
R
Am/V
Experimental
DR AS/NZS 2327
mins
m− 1
Ts o C
Tc o C
Tr o C
Ts o C
Tc o C
Tr o C
46 23 72
20 20 18
770 676 929
– 172 –
– 156 –
800 633 928
476 320 581
376 185 529
c) Fibre reinforced concrete Specimen
F1 F2 F5 F6
R
Am/V
Experimental
DR AS/NZS 2327
mins
m− 1
Ts o C
Tc o C
Ts o C
Tc o C
24 25 85 51
20 20 18 18
666 647 963 693
292 214 – 395
641 650 972 825
328 335 637 474
Note: “–” means un-reliable data. Note: “-” means un-reliable data. Table 5 Comparison of experimental fire axial loads with values calculated from DR AS/NZS 2327 and EN 1994-1-2. a) Plain concrete Specimen
DR AS/NZS 2327
R
Eƒi,Exp
mins
kN
37 26 38
1378 1378 1415
0.7 0.85 0.7
b) Conventionally reinforced concrete R1 46 1485 R2 23 1485 R5 72 1604 c) Fibre reinforced concrete F1 24 F2 25 F5 85 F6 51
Eƒi,Rd kN
Eƒi,Exp
1.29 1.44 1.10
1311 1312 1917
1.05 1.05 0.74
0.5 0.7 0.5
0.7 0.85 0.7
1.36 1.42 1.39
1180 1723 1042
1.26 0.86 1.53
0.7 0.85 0.7 0.85
1.18 1.43 1.14 1.14
1968 1478 760 1167
0.5 0.93 1.86 1.21
k
P1 P2 P5
1378 1378 1415 1415
λT
EN 1994-1-2 k
λT
Eƒi,Rd kN
Eƒi,Exp
0.94 1.21 0.80
2068 1809 2699
0.66 0.76 0.52
0.5 0.7 0.5
0.98 1.19 1.01
1927 2380 1892
0.77 0.62 0.84
0.5 0.7 0.5 0.5
0.83 1.19 0.98 1.29
2895 2040 1095 1630
0.47 0.67 1.29 0.87
Eƒi,Rd
Eƒi,Rd
0.49 are conservative when the “k” values as specified in DR AS/NZS 2327 are applied. Therefore, the following conclusion can be drawn:
a measured concrete compressive strength ranging from 80 MPa to 95 MPa. All tests were conducted on square hollow steel sections having nominal cross-sections dimensions of 200 mm × 200 mm × 6 mm and 220 mm × 220 mm × 6 mm; the yield strengths of the sections tested ranged from 461 MPa to 569 MPa. In the tests, end restraint conditions of fixed-fixed (F-F) and pinned-fixed (P-F)) with considered. It was found that CFST columns filled with conventional reinforced concrete having a fixed-fixed boundary condition and columns filled with steel fibre reinforced concrete having steel contribution ratio of 0.47 and
1. A comparison of the experimental axial capacity in fire with that determined by calculation to DR AS/NZS 2327 and EN 1994-1-2 has shown that the standards can be un-conservative, particularly for columns filled with plain concrete and with an R value of < 30 min. For these cases, an effective length factor value of 0.9 has been shown to be more representative for continuous CFST column where
108
Journal of Constructional Steel Research 136 (2017) 101–109
K.U. Ukanwa et al.
References
Table 6 Comparison of experimental fire axial load using “k” factor of 0.9.
[1] M. Yu, X. Zha, J. Ye, B. Wang, A unified method for calculating fire resistance of solid and hollow concrete-filled steel tube columns based on average temperature, Eng.Struct. 71 (2014) 12–22. [2] ISO I. 834, Fire Resistance Tests-elements of Building Construction, International Organization for Standardization, Geneva, Switzerland, 1999. [3] M.L. Romero, V. Moliner, A. Espinos, C. Ibañez, A. Hospitaler, Fire behavior of axially loaded slender high strength concrete-filled tubular columns, J. Constr. Steel Res. 67 (12) (2011) 1953–1965. [4] V.K.R. Kodur, J. Latour, Experimental Studies on the Fire Resistance of Hollow Steel Columns Filled With High-strength Concrete, (2005). [5] T. Lie, R. Irwin, Fire resistance of rectangular steel columns filled with barreinforced concrete, J.Struct.Eng 121 (5) (1995) 797–805. [6] N. Mago, S.J. Hicks, Fire behaviour of slender, highly utilized, eccentrically loaded concrete filled tubular columns, J. Constr. Steel Res. 119 (2016) 123–132. [7] A. Espinos, M.L. Romero, A. Hospitaler, Advanced model for predicting the fire response of concrete filled tubular columns, J. Constr. Steel Res. 66 (8) (2010) 1030–1046. [8] Z. Wang, J. Wu, J. Wang, Experimental and numerical analysis on effect of fibre aspect ratio on mechanical properties of SRFC, Constr. Build. Mater. 24 (4) (2010) 559–565. [9] P. Schaumann, V. Kodur, O. Bahr, Fire behaviour of hollow structural section steel columns filled with high strength concrete, J. Constr. Steel Res. 65 (8) (2009) 1794–1802. [10] V. Moliner, A. Espinos, M.L. Romero, A. Hospitaler, Fire behavior of eccentrically loaded slender high strength concrete-filled tubular columns, J. Constr. Steel Res. 83 (2013) 137–146. [11] V.K. Kodur, R. Fike, Response of concrete-filled HSS columns in real fires, Eng. J. AISC Inc. 46 (4) (2009) 243–256. [12] A. Espinos, M.L. Romero, E. Serra, A. Hospitaler, Experimental investigation on the fire behaviour of rectangular and elliptical slender concrete-filled tubular columns, Thin-Walled Struct. 93 (2015) 137–148. [13] A. Espinos, M.L. Romero, E. Serra, A. Hospitaler, Circular and square slender concrete-filled tubular columns under large eccentricities and fire, J. Constr. Steel Res. 110 (2015) 90–100. [14] DR AS/NZS 2327, Composite Steel-concrete Construction for Buildings, Australian/ New Zealand Standards, 2016. [15] A. Espinos, M.L. Romero, A. Hospitaler, Simple calculation model for evaluating the fire resistance of unreinforced concrete filled tubular columns, Eng. Struct. 42 (2012) 231–244. [16] EN B. 1363–1, Fire Resistance Tests. General Requirements. BSI Standard 2012, (2012). [17] Eurocode E, 4: Design of composite steel and concrete structures–part 1–2: general rules–structural fire design, British Standards Institution, BS EN, 1994, pp. 1–2. [18] British Standards Institution, BS 8666: 2005: scheduling, dimensioning, bending and cutting of steel reinforcement for concrete, Specification: British Standards Institution, 2005. [19] M. Chabot, T.T. Lie, Experimental Studies on the Fire Resistance of Hollow Steel Columns Filled with Bar-reinforced Concrete, (1992). [20] J. Ding, Y. Wang, Realistic modelling of thermal and structural behaviour of unprotected concrete filled tubular columns in fire, J. Constr. Steel Res. 64 (10) (2008) 1086–1102. [21] S. Hong, A.H. Varma, Analytical modeling of the standard fire behavior of loaded CFT columns, J. Constr. Steel Res. 65 (1) (2009) 54–69. [22] K. Chung, S. Park, S. Choi, Fire resistance of concrete filled square steel tube columns subjected to eccentric axial load, Int. J. Steel Struct. 9 (1) (2009) 69–76. [23] J. Yin, X. Zha, L. Li, Fire resistance of axially loaded concrete filled steel tube columns, J. Constr. Steel Res. 62 (7) (2006) 723–729. [24] M. Spearpoint (Ed.), Fire Engineering Design Guide, third ed., Centre for Advanced Engineering: Christchurch, New Zealand, 2008.
a) Plain concrete Specimen
R
Eƒi,Exp
mins
kN
37 26 38
1378 1378 1415
DR AS/NZS 2327 λT
EN 1994-1-2
Eƒi,Rd kN
Eƒi,Exp
889 1199 1370
1.55 1.15 1.07
b) Conventionally reinforced concrete R1 46 1485 1.36 1180 R2 23 1485 1.49 1577 R5 72 1604 1.39 1042 c) Fibre reinforced concrete F1 24 1378 F2 25 1378 F5 85 1415 F6 51 1415
P1 P2 P5
1.63 1.51 1.39
1.48 1.50 1.14 1.44
1389 1351 760 1166
λT
Eƒi,Rd kN
Eƒi,Exp
1.63 1.51 1.39
919 1243 1373
1.50 1.11 1.03
1.26 0.94 1.54
1.36 1.49 1.39
1232 1636 1193
1.2 0.91 1.34
0.99 1.02 1.86 1.21
1.48 1.50 1.14 1.14
1440 1402 806 1208
0.96 0.98 1.75 1.17
Eƒi,Rd
Eƒi,Rd
the member temperatures are calculated in accordance to DR AS/ NZS 2327; based on the experimentally observed effective length of these columns. This effective length factor was suggested from examination of the lateral deformed shape of the tested specimens. 2. For columns having slenderness ranging from 0.47 to 0.67 in the ambient temperature condition, and with percentage of bar reinforcement in accordance to EN 1994-1-2, it was found that columns filled with steel fibre reinforced concrete had the highest structural fire resistance to the concentric compression loaded columns. This was due to the even distribution of steel fibre in the concrete core, which increased the tensile strength of the concrete core and so increased the bending resistance. 3. The longitudinal elongation of the steel was < 3 mm. This is much less than the longitudinal elongation recorded in other CFST fire tests where the entire height was heated. This was as a result of the shear bond between the steel and concrete at the cooler top and bottom region of the steel which restrained the heated part from further elongation.
Acknowledgements The authors wish to express gratitude to the New Zealand Heavy Engineering Educational & Research Foundation (HEERF) for their scholarship support, also to the department of Civil Engineering, IIT Roorkee for providing the laboratory used for the experiments.
109