Full-scale fire resistance tests of lightweight steel framed floor systems

Thin–Walled Structures 192 (2023) 111117 Available online 6 September 2023 0263-8231/© 2023 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/bync-nd/4.0/). Full length article Full-scale fire resistance tests of lightweight steel framed floor systems Martin Hor´ aˇ cek a , * , Thomas Gernay b , Marcela Karmazínov´ a a , Marek Hladík c , Zbynˇ ek P¨ offel c , Ondˇ rej Peˇ sek a , Ivan Bal´ azs a a Institute of Metal and Timber Structures, Brno University of Technology - Faculty of Civil Engineering, Veveˇ rí St. 95/331, Brno, 60200, Czech Republic b Department of Civil and Systems Engineering, Johns Hopkins University – Whiting School Engineering, 3400 North Charles Street, Baltimore, MD 21218, United States c voestalpine Profilform s.r.o., Tov´ arní St. 4, Vyˇ skov, 68223, Czech Republic ARTICLE INFO Keywords: Cold-formed steel section Full-scale test Fire resistance Lipped channel beam LSF floor systems ABSTRACT As cold-formed steel profiles are increasingly used for load-bearing structural systems, their fire design requires specific attention due to their thin-walled nature and high slenderness. In some specific applications such as in storage hall floor systems, these structures are used without sheathing or thermal protection on the fire-exposed side. Yet, there is a lack of data from fire resistance tests on full-scale load-bearing thin-walled steel structures, especially with directly exposed steel. This article describes two standard fire resistance tests on full-scale light gauge steel frame floors made of cold-formed steel lipped channel girders and joists topped by chipboard panels. The experimental program was designed to investigate the fire resistance of the unprotected girders. A specificity of this program was that the girders were subjected to a low load level to probe the ability to achieve without passive fire protection a 30-min fire resistance rating typical for storage hall structures in the Czech Republic. The absence of protection resulted in differences in thermal gradients and bracing compared to common fire tests, as well as a very low degree of utilization leading to an expected failure temperature higher than 800 ◦C. The results showed that the two floors remained stable during the 30 min with limited deflections, but failed the deflection rate criteria after 24 and 22 min, respectively. Comparison is provided with the calculation methods from the Eurocodes. The presented results provide new data on the response of full-scale, unprotected coldformed steel floor systems subjected to fire, which can be used to calibrate numerical models and fire design methods outside of the range currently covered by the codes. 1. Introduction Thin-walled cold-formed steel (CFS) members are used in civil engineering for various applications, such as structural members of residential and industrial buildings, floor systems, and storage racks [1]. Systems of thin-walled beams include purlins, wall girts, frame rungs, girders, and joists, while columns include frame struts or uprights in racking systems, with a range of cross-sectional shapes. Their structural design requires a thorough understanding of the stability behavior of the members, including global, local, and distortional buckling, and their interactions. While methods have been developed for determining the design strength under normal conditions (i.e., effective width method, direct strength method), data and knowledge on the behavior at elevated temperatures are limited. In a fire situation, the steel’s mechanical properties are reduced with increasing temperature. Previous research has focused on hot-rolled steel [2], but the properties of coldformed steel at elevated temperatures may differ [3] and the provisions in the current European Standard for structural fire design [4] do not distinguish between hot-rolled and cold-formed steel, despite indications that cold-formed steel may be more sensitive to elevated temperatures [5]. Besides, unlike most hot-rolled members, cold-formed steel members commonly use non-doubly symmetrical shapes and have their strength influenced by local and distortional buckling, which renders necessary the specific study of cold-formed steel members in the fire. Several studies have investigated the material properties and behavior of cold-formed steel members under fire situations, including lateral–torsional buckling resistance, the behavior of wall and floor systems, and the effects of insulation and connections on fire resistance. Arrais et al. [6] studied the lateral–torsional buckling resistance of coldformed steel beams of channel sections at elevated temperatures using numerical simulations. The analyses indicated the conservativeness of * Corresponding author. E-mail address: [email protected] (M. Hor´ aˇ cek). Contents lists available at ScienceDirect Thin-Walled Structures journal homepage: www.elsevier.com/locate/tws https://doi.org/10.1016/j.tws.2023.111117 Received 14 April 2023; Received in revised form 14 July 2023; Accepted 12 August 2023 Thin-Walled Structures 192 (2023) 111117 2 the provisions for design buckling resistance determination in the European standard and proposed some modifications of the calculation methods. Numerical models calibrated based on fire experiments were used by Laím and Rodrigues [7] which resulted in the proposal of a new methodology for the determination of design buckling resistance of coldformed steel isolated beams at elevated temperatures. It was emphasized that provisions in the European standard are not fully suitable for coldformed members. Gunalan et al. [8] performed fire experiments on wall systems made of steel framing of thin-walled cold-formed members of channel cross-sections with various types of insulations. Other studies of similar structures also considered plasterboards on both sides of the steel cold-formed wall framing systems [9,10]. Other studies investigated the post-fire local buckling of circular hollow sections [11], the fire resistance of cold-formed box columns [12] and built-up columns [13], and the behavior of individual cold-formed steel beams in fire [14,15]. Highstrength cold-formed steel members have also been studied [16–18]. Floor systems consisting of cold-formed thin-walled members of channel sections were numerically and experimentally investigated by Baleshan and Mahendran [19,20]. In their studies, the steel framing was on both sides covered by plasterboards or plywood, respectively, which contributed to the higher fire resistance of the steel members due to reducing their temperature during a fire in time and also ensured lateral restraint of the members. Steau and Mahendran [21] performed fire resistance tests of relatively small specimens consisting of steel framing protected with plasterboards and steel sheathing and evaluated the contribution of the protection to the resistance. Another area of study is the behavior and possible failure of connection of the steel members and floor to the lining (e.g., plasterboards). Screw connections are usually used. It was shown [22] that failure of the connection can directly affect the resistance of the members due to the reduction of the restraint provided to the steel member by adjacent planar members of coverings. However, significant knowledge gaps remain in understanding the behavior and fire resistance of thin-walled CFS members. While the above studies provided very valuable insights, the literature review reveals that full-scale fire experiments have rarely been conducted, therefore, observations on the behavior of the entire CFS systems are lacking. Furthermore, previous studies suggest that the provisions in the Eurocode may not be fully suitable for the fire design of CFS beams and channels [6,7]. This warrants further research and shows the need for additional test data to calibrate models for floor systems, and eventually support the development of applicable design provisions. Finally, previous experimental studies on CFS systems have largely focused on assemblies with fire protection (i.e., sheathing). This is because in practice the CFS members are most often protected from fire with plasterboard or plywood. The presence of this sheathing influences the stability behavior of the CFS members through bracing. However, in some specific applications, CFS floor systems may be constructed without passive fire protection. To the authors’ knowledge, no full-scale loaded fire test of such assembly has previously been published. While common fire design solutions for CFS systems rely on passive protection [23], there are applications where passive fire protection is either impossible or undesirable, for instance, due to economic or aesthetic reasons. One such example are lightweight steel framed (LSF) built-in floor systems for storage halls (Fig. 1). These systems have seen increased demand in recent years due to the growing e-commerce industry. They represent an effective solution for expanding warehouse space in logistics halls. They can be used from small applications of a few tens of square meters to storage floors of several thousand square meters. The floor structure consists of an arrangement of primary beams and joists supporting chipboard panels. The beams and joists are made of cold-formed steel lipped channel sections. Connecting angles with fasteners are used. For economic reasons, there is an interest in using these floors without fire protection. The requirement of 30 min of fire resistance can then only be achieved if the load level on the structure is very low, but for such structures, this may be practical because the selfweight is low and the load combination in the fire situation uses low coefficients for other accompanying actions. The absence of protection leads to differences in thermal gradients and bracing compared to protected assemblies. Yet, there is a lack of data from fire resistance tests on such unprotected CFS systems. To address this gap, this article describes two full-scale fire resistance tests on sections of the floor structure used for built-in floors for storage halls. The two experimental tests described in this paper were conducted according to the fire resistance test standards [24,25]. The experimental results are described, and a comparison is conducted between the observed behavior and the fire design provisions from the European standards (Eurocodes). The Eurocode EN 1993-1-2 [4] provides methods for the determination of the fire resistance by calculation of steel members with Class 4 cross-sections in Annex E. These provisions also apply to cold-formed sections. The strength determination requires using the effective section characteristics (i.e., effective cross-section area and effective crosssection modulus in accordance with EN 1993-1-3 [26] and EN 1993-1-5 [27]) with the value of the elevated temperature yield strength at 0.2% proof stress (i.e., using Table E.1 and Fig. E.2 in lieu of Table 3.1 and Fig. 3.1 of the standard) [28]. The Czech National Annex of the standard [4] further provides (see cl. 4.2.3.6) an equation for the critical temperature of members with 4 Class cross-sections as a function of the degree of utilization, or alternatively states that the critical temperature for these members may be considered as 500 ◦C and 450 ◦C in bending and compression, respectively. Therefore, one of the objectives of this paper is to compare the results of the experimental tests with these provisions, to assess their applicability for the considered unprotected cold-formed steel members. It is expected that, for the thin-walled steel members directly exposed to fire, the steel temperature at 30 min will be very high. However, the tests will allow probing whether the structure can still achieve fire resistance under low load ratio, and hence provide new data points for assessing the applicability of the critical temperature method for such assemblies. 2. Description of the experimental program The experimental program was focused on the determination of the fire resistance of the primary beams. Two different arrangements of built-in LSF floor system sections were subjected to experimental investigation labeled fire test configurations A and B. The assemblies of both configurations before the installation of the chipboard panels are shown in Fig. 2. The design of both assemblies was conducted based on the rules in application in the Czech Republic. Specifically, design at ambient temperature was conducted according to EN 1993-1-1 [29] and EN 1993-1-3 [26]. The governing failure mode is the failure of the primary beams in bending. Lateral–torsional buckling of the primary beams is prevented through the continuous connection between the top flanges and the chipboard panels. As the primary beams are the focus of the study, the joists and connections were designed to withstand fire exposure. For the calculation of the effects of loads at ambient temperature, the load combination (6.10) according to EN 1990 [30] was applied. The permanent loads (Gk) are the self-weight of the structural members of the floor system, including primary beams, joists, angle cleats, and chipboard panels, equal to 0.25 kN/m 2 . The load factor γG is equal to 1.35. The imposed loads (Qk) were selected to obtain a design fire resistance of 30 min for the tested assemblies. Accordingly, Qk was 2.10 kN/m 2 for the test specimen of configuration A and 1.90 kN/m 2 for the test specimen of configuration B. The corresponding values of imposed loads for the fire situation were applied in the fire tests using steel ballast evenly distributed over the top surface of the chipboard panels. It is noted that the studied LSF floor systems fall into category E1, i.e., areas susceptible to accumulation of goods, including access areas. In the Czech Republic, the recommended value of surface load qk for category E1 is 7.50 kN/m 2 , but the values may be changed according to the usage of the particular project. The load factor γQ is 1.50. M. Hor´ aˇ cek et al. Thin-Walled Structures 192 (2023) 111117 3 The fire load combination (accidental design situation) was calculated according to EN 1991-1-2 [31], cl. 4.3.1; with ψ 2,1=0.8 (cat. E). The reduction factor η fi for load combination (6.10) in EN 1990 was determined according to EN 1993-1-2, cl. 2.4.2; Eq. (2.5). Table 1 summarizes the applied loading on the specimens. With the variable load Qk equal to 2.10 kN/m 2 and 1.90 kN/m 2 for specimens A and B, respectively, the fire load combination yields an applied load on the specimens during the fire tests of 1.68 kN/m 2 and 1.52 kN/m 2 , respectively. 2.1. Set-up of fire test configuration A In the fire test configuration A (Fig. 3), the tested section of the floor structure consisted of one internal double primary beam (two single profiles screwed together with M16 8.8 bolts over insert plates of a thickness of 10 mm — in Fig. 3c as A.1) and two outer single primary beams (in Fig. 3c as A.2), both with a length of 5060 mm and span of 5000 mm (distance between the bolted connection of primary beams to gusset plated welded on columns). The profiles of the primary beams (both double internal and outer single) are thin-walled cold-formed channel sections with double-edge flange stiffeners with a profile depth of 402 mm, flange width of 110 mm, and a plate thickness of 3.2 mm (see Fig. 4a). The manufacturer was voestalpine Profilform s.r.o., with profile designation METSEC 402 C+32 produced from hot-dip galvanized steel S450GD (actual values: Reh =452 MPa, Rm=545 MPa). The axial distance between the internal and outer primary beams is 1653 mm. In total 2 ×6 =12 joists (in Fig. 3c as A.3) of length of 1420 mm Fig. 1. Storage hall with a built-in floor system. Fig. 2. Specimens of fire test configuration A (left top) and fire test configuration B (right bottom) shown before installation of the chipboard panels. Table 1 Load combinations for configuration A and configuration B. Load combinations Config. A Config. B Permanent 0.25 kN/ m 2 0.25 kN/ m 2 Variable 2.1 kN/ m 2 1.9 kN/ m 2 Load combination (6.10) acc. EN 1990; γG =1.35; γQ,1 = 1.5 3.49 kN/ m 2 3.19 kN/ m 2 Fire combination acc. EN 1991-1-2, cl. 4.3.1; ψ 2,1 =0.8 (cat. E) 1.93 kN/ m2 1.77 kN/ m 2 Applied load on the specimen Fire combination load excluding permanent load 1.68 kN/ m 2 1.52 kN/ m 2 Reduction factor η fi acc. EN 1993-1-2, cl. 2.4.2; Eq. (2.5) 0.553 0.555 M. Hor´ aˇ cek et al. Thin-Walled Structures 192 (2023) 111117 4 were used in test configuration A. The axial distance between the joists is 840 mm. The profiles of the joists are thin-walled cold-formed channel sections with single-edge flange stiffeners with a profile depth of 342 mm, flange width of 100 mm, and a plate thickness of 2.3 mm (profile designation METSEC 342 M 23 — see Fig. 4b), steel grade S450GD (actual values: Rp,02 =455 MPa, Reh =472 MPa, Rm=565 MPa). The joists are connected to the primary beams with M16 8.8 bolts via angle cleats (thickness of 6 mm, steel grade S235 — see Fig. 4c). The primary beams are supported by 6 short columns (in Fig. 3c as A.4) of length 550 mm. The columns consist of two steel hot-rolled U profiles with a depth of 100 mm, orientated back to back, keeping a gap of 10 mm between U profiles. At the top end of the column, a gusset plate Fig. 3. Scheme of fire test configuration A. Fig. 4. Dimensions of beam profiles and angle cleats. M. Hor´ aˇ cek et al. Thin-Walled Structures 192 (2023) 111117 5 of thickness 10 mm is inserted into the gap and welded to the U profiles. The primary beams are connected to gusset plates with two M16 8.8 bolts at each primary beam’s end. On the bottom end of the column, the steel base plate of dimensions 200 ×200 and thickness of 15 mm is welded to the U profiles. As a floor deck, 6 pieces of Flameshield Supreme Deluxe floor chipboard panels, manufactured by BERGHOEF GmbH in Germany, of a thickness of 38 mm are used (volume weight of 711 kg/m 3 , modulus of elasticity of 2600 MPa, P6 — bending strength 15 MPa). Fire protection classification of chipboards panels according to EN 13501 is for top surface Blf-s1, respectively for bottom surface Blfs1/d0 (Bfl =flame-retardant construction products — radiation intensity of 8 kW/m 2 ; s1 =no smoke production; d0 =no droplets). The dimensions of chipboard panels are 1680 ×1655 mm, 1385 ×1655 mm, and 2225 ×1655 mm. The chipboard panels are anchored to the joists using WINGTEKS 6.3 ×70 mm screws, manufactured by ETANCO in France, at a spacing of 400 to 600 mm. Fig. 5 shows the test speciment of configuration A from the heated side. 2.2. Set-up of fire test configuration B In the fire test configuration B (Fig. 6) the specimen was made of two outer single primary beams (in Fig. 6c as B.1) of 5060 mm length and 5000 mm span The axial distance between the outer primary beams is 2000 mm. The six joists (in Fig. 6c as B.2) are 1780 mm long with an axial distance of 840 mm between them. The same profiles of primary beams (METSEC 402 C+32) and secondary beams (METSEC 342 M 23) were used as in the test configuration A. Connections between the joists and the primary beams, and between the primary beams and the four short columns (in Fig. 6c as B.3), are similar to those of specimen A. For the floor deck, 6 pieces of the same Flameshield Supreme Deluxe floor chipboards as in specimen A are used. Unlike in fire test configuration A, the joints between chipboard panels are covered with 200 mm wide blanks made of chipboard of thickness 20 mm, screwed to the chipboard panels with 6 ×50 mm screws at a spacing of 600 mm. Fig. 7 shows the test speciment of configuration B from the heated side. 2.3. Tests preparation After the assembly, the test specimens were seated on the test furnace by crane. Steel columns were spot welded to the rigid thresholds of the test furnace (Fig. 8), with insulation of exposed parts of the profiles with mineral wool. The sides of the specimen were left free for the test, without restricting the deflection freedom of the loaded structure. The specimens were loaded from above using steel ballasts simulating a uniform continuous load with an intensity based on values listed in Table 1 (see Fig. 9 and Fig. 10 for test configuration A and test configuration B, respectively). 2.4. Temperature measurements The test furnace was heated by a system of diesel burners. The furnace temperature was measured by plate temperature sensors and recorded at minute intervals. The measuring ends of the plate temperature sensors were evenly spaced 100 mm from the exposed surface of the sample (temperature sensors are visible in Fig. 7) according to Article 9.1.1 [25]. The temperatures in the furnace were regulated so that within the prescribed tolerances (see [24] Article 5.1.2) they corresponded to the relationship according to [24] Article 5.1.1. The overpressure in the test furnace was measured and regulated so that the values corresponded to the conditions [24] of Article 5.2.1. Temperatures on the unheated surface of the sample were measured by K-type disk thermometric cells and recorded at minute intervals. Measuring joints of thermometric cells soldered to the center of a copper target with a diameter of 12 mm and a thickness of 0.2 mm and covered with a 2 mm thick plate measuring 30 ×30 mm (see [24] Article 4.5.1.2), mounted on the surface of the sample according to [25] Articles 9.1.2.2 and 9.1.2.3. The temperatures were measured also inside the sample, namely, between chipboard panels and primary or secondary beams, and on the bottom flange of the primary or secondary beams. Fig. 11 provides the overview of installed sensors on the test specimen of configuration A. The marks designated with numbers 20 to 24 (see left part of Fig. 11) represent the positions of temperature sensors installed on the top (unheated) surface of chipboard panels. The marks designated with numbers 25 to 38 (see the middle part of Fig. 11) represent the positions of temperature sensors installed on the primary beams and joists. In each position in the layout of the test specimen, two temperature sensors are installed. The blue numbers are for sensors installed on the top surface of the bottom flanges of beams, while the red numbers are for sensors installed on the top surface of the top flanges of beams (sensors are installed in between the beam top flange and chipboard panels). Correspondingly, Fig. 12 provides the overview of installed sensors on the test specimen of configuration B. During the testing, the deflections of primary beams at their midspan were measured. In total 4 deflection gauges were installed in test configuration A (designated as D1 to D4, see Fig. 11c), and 2 deflection gauges in test configuration B (designated as D1 and D2, see Fig. 12c). Fig. 5. Fire test configuration A – view of the heated side of the specimen. M. Hor´ aˇ cek et al. Thin-Walled Structures 192 (2023) 111117 6 2.5. Tests procedure The fire resistance tests were performed according to EN 1365-2 Fire resistance tests for loadbearing elements — Part 2: Floors and roofs [25]. The tests were carried out for 30 min, after which the burners were turned off. Failure of the floor system was defined as the time when the ability to support the test load was lost, as defined in EN 1363-1 [24]. Specifically, failure was deemed to have occurred when either the deflection or the rate of deflection exceeded the threshold specified in the fire resistance test standard. Eq. (1) specifies the threshold for deflection, and Eq. (2) the threshold for the rate of deflection, both with substituted values valid for tested configurations. Fig. 6. Scheme of fire test configuration B. Fig. 7. Fire test configuration B – view of the heated side of the specimen. M. Hor´ aˇ cek et al. Thin-Walled Structures 192 (2023) 111117 7 For the deflection: D=L2 400⋅d=50002 400⋅402 =155.5 mm (1) For the rate of deflection: dD dt =L2 9000⋅d=50002 9000⋅402 =6.9 mm/min (2) where D is the limiting deflection (in mm), L is the clear span of the test specimen (in mm), d is the distance from the extreme fiber of the cold design compression zone to the extreme fiber of the cold design tension zone of the structural section (in mm), and dD/dt is the limiting rate of Fig. 8. Installed specimen; columns welded to the furnace thresholds at both ends of the furnace. Fig. 9. Fire test configuration A – specimen loaded from above with steel ballasts. Fig. 10. Fire test configuration B – specimen loaded from above using steel ballasts. M. Hor´ aˇ cek et al. Thin-Walled Structures 192 (2023) 111117 8 deflection (in mm/min) [32]. 3. Fire resistance calculation according to Eurocode The assessment of structures for the effects of fire can be performed either in the temperature domain (critical temperature) or the resistance domain by comparing the load-bearing capacity at elevated temperatures with the effects of loads calculated for a fire load combination. The steel members were not protected against fire. For these storage hall structures, if fire resistance is required, in the Czech Republic the usual fire resistance rating is 15 min, with a maximum of 30 min. However, it is common for these structures to not require fire resistance if active protection measures such as sprinklers are in place. There is also a strong financial incentive in achieving this rating with lightweight steel framed floor structural systems without the use of fire protection on the steel members. Therefore, the sizing of the steel members is carried out to achieve 30 min of standard fire resistance, which is the maximum requirement in practice, with the unprotected profiles using the Eurocode EN 1993-1-2 [4] calculation method. Specifically, the steel profile is chosen such that its critical temperature under the relevant load combination in the fire situation is higher than the temperature reached after 30 min of fire exposure of the unprotected profile. This leads to a low utilization ratio of the load-bearing elements at ambient temperature (in time t =0), with a degree of utilization μ 0 acc. to EN 1993-1-2, cl. 4.2.4, Eq. (4.23) equal to 0.059 and 0.069 for the internal primary beam of specimen A and the outer primary beams of specimen B, respectively. Nevertheless, for this type of structure, it may remain economical to overdesign the thin-walled members at ambient temperature to be able to omit the fire protection. For the primary beams under study, the critical temperature approach can be used since their load-bearing capacity is proportional to the effective yield strength (with proper consideration of local buckling of the plates for the class 4 cross-sections), as lateral–torsional buckling of the beams and joists is prevented through the continuous connection between the top flanges and the chipboard. For hot rolled sections, the critical steel temperature can be determined according to EN 1993-1-2 [4], cl. 4.2.4, Eq. (4.22). This equation is applicable for a degree of utilization not smaller than 0.013. The Czech National Annex NA.2.5 of the standard [4] specifies the expression for the critical steel temperature calculation for the design of members with class 4 cross-section with the design yield strength reduced according to Table E.1 [4], see Eq. (3): θa,cr =36.5⋅ln[1 1.0⋅ μ 4,167 0 −1]+435 (3) Fig. 11. Measurements position in test configuration A: the temperature on the unheated side (a); the temperature of steel members (b); the deflections of primary beams (c). Fig. 12. Measurements position in test configuration B: the temperature on the unheated side (a); the temperature of steel members (b); the deflections of primary beams (c). M. Hor´ aˇ cek et al. Thin-Walled Structures 192 (2023) 111117 9 The condition for the fire design is fulfilled provided the steel member temperature θ a calculated at time t (taken as the required fire resistance) is less than the critical temperature θ a,cr. Alternatively, the fire resistance can be determined based on the load-bearing capacity of the steel member for the elevated temperature θ a at time t. The loadbearing capacity directly depends on the effective yield strength fy,θ, which is calculated for the elevated temperature using reduction factor ky,θ according to EN 1993-1-2 [4], chap. 3. With the reduced yield strength, the corresponding design load-bearing capacity Rfi,d,t is calculated and compared with the design load effect at the fire situation Efi,d according to EN 1991-1-2 [31]. The temperature increase in the exposed steel profiles is evaluated using the method from Eurocode EN 1991-1-2 [31], chap. 3 +EN 19931-2 [4], cl. 4.2.5 (Eq. 4.25). The method is based on a lumped mass approach. The profiles are unprotected. The determination of Eurocode design fire resistance for both tested configurations based on the design in the temperature domain (critical temperature) is summarized in Table 2 and Table 3 for secondary and primary beams, respectively. The effective cross-section properties taking into account the effects of local buckling of compressed slender section parts are used for the calculation of the bending resistance at ambient temperature (note: at the time of preparation of the fire resistance tests, the effects of distortional buckling were not considered in the design fire resistance calculations). When calculating the correction factor for the shadow effect ksh, the 3sided fire is considered since the chipboard panels protect the profile of primary or secondary beams against the fire from the top side. In addition, in the case of double inner primary beams in test configuration A, shading of the inner surfaces of the webs is considered (fire affects only one surface of the web of the profiles, the surface of the web on the other side is shaded by the web of the other profile). In the case of secondary beams, the critical temperature is reached after an hour (62.2 min and 55.6 min for specimens A and B, respectively). Thus, secondary beams reach the critical temperature after about double the time in comparison with the required fire resistance. However, it is necessary to note, that secondary beams have been overdimensioned from the fire resistance viewpoint because the main aim was to find out the fire resistance of primary beams as the main beams of the structural system. The primary beams achieve the 30-min fire resistance rating according to these calculation methods from the Eurocode. The calculated time of failure is 36 min and 30 min for specimens A and B, respectively. Alternatively, the determination of the Eurocode design fire resistance of primary beams considering the effective cross-section properties taking into account the effects of local and distortional buckling is provided in Table 4. The effect of distortional buckling is reflected in the reduction of the effective cross-sectional characteristics (sectional modulus), which leads to reduced bending resistance at ambient temperature and a higher degree of utilization at the same load. The increase in the degree of utilization results in a decrease in the critical temperature. The steel temperature is not affected by the change of effective cross-sectional properties since the mass and surface of the primary beams profile exposed to the fire remain the same. The lower critical temperature would occur earlier (see Tables 3 and 4 - last rows), so that the test results – in comparison with this calculated critical temperature and the time of its reaching – give some, even a small reserve. The consideration of the effective cross-section characteristics including distortional (not only local) buckling influence will contribute to more accurate calculation and will help to more significant comparison with the test results and the verification of the calculating procedure. It also shows, each influence, apparently less significant, can influence the fire resistance by a few minutes, which can be important for the structural design, especially in the case of steel structural members with generally small fire resistance. 4. Results of fire tests 4.1. Configuration A The measured temperatures in the furnace are plotted in Fig. 13 (the red dashed lines for each temperature sensor). The average value of temperature is presented by the blue solid line. The temperature in the furnace shows a good match with the standard temperature curve Table 2 Determination of Eurocode design fire resistance of secondary beams taking into account the effects of local buckling. Secondary beams Config. A Config. B Length 1.65 m 2.0 m Profile 342M23 342M23 Section factor Am/V 823.2 m −1 823.2 m −1 Box value of section factor [Am/V]b 623.1 m −1 623.1 m −1 Shear force Vfi,Ed 1.42 kN 1.58 kN Moment Mfi,Ed 0.58 kNm 0.79 kNm Resistance at ambient temperature (at time t=0) Rfi,d,0(shear is decisive) 41.81 kN 41.81 kN Degree of utilization μ 0at time t=0 acc. to EN 1993-1-2, cl. 4.2.4, Eq. (4.23) 0.034 0.038 Steel temperature (30 min) θa acc. EN 1991-1-2, chap. 3 +EN 1993-1-2, cl. 4.2.5 839 ◦C 839 ◦C Critical temperature θa,cr acc. EN 1991-1-2, cl. 4.2.4; Eq. (4.22) 950 ◦C 933 ◦C Critical temperature reached acc. EN 1993-1-2, cl. 4.2.5.1; Eq. (4.25) 62.2 min 55.6 min Table 3 Determination of Eurocode design fire resistance of primary beams taking into account the effects of local buckling. Primary beams Config. A Config. B Length 5.0 m 5.0 M Profile 402Cþ32 402Cþ32 Section factor Am/V 406.8 m −1 592.7 m −1 Box value of section factor [Am/V]b 236.7 m −1 422.5 m −1 Shear force Vfi,Ed 9.30 kN 5.45 kN Moment Mfi,Ed 11.60 kNm 6.80 kNm Resistance at ambient temperature (at time t=0) Rfi,d,0(bending is decisive) 196.4 kNm 98.2 kNm Degree of utilization μ 0at time t=0 acc. to EN 1993-1-2, cl. 4.2.4, Eq. (4.23) 0.059 0.069 Steel temperature (30 min) θa acc. EN 1991-1-2, chap. 3 + EN 1993-1-2, cl. 4.2.5 830 ◦C 837 ◦C Critical temperature θa,cr acc. EN 1991-1-2, cl. 4.2.4; Eq. (4.22) 865 ◦C 841 ◦C Critical temperature reached acc. EN 1993-1-2, cl. 4.2.5.1; Eq. (4.25) 36.7 min 30.7 min Table 4 Determination of Eurocode design fire resistance of primary beams taking into account the effects of local and distortional buckling. Primary beams Config. A Config. B Profile 402Cþ32 402Cþ32 Resistance at ambient temperature (at time t=0) Rfi,d,0(bending is decisive) 171.94 kNm 85.97 kNm Degree of utilization μ 0at time t=0 acc. to EN 1993-1-2, cl. 4.2.4, Eq. (4.23) 0.067 0.079 Steel temperature (30 min) θa acc. EN 1991-1-2, chap. 3 + EN 1993-1-2, cl. 4.2.5 830 ◦C 837 ◦C Critical temperature θa,cr acc. EN 1991-1-2, cl. 4.2.4; Eq. (4.22) 845 ◦C 821 ◦C Critical temperature reached acc. EN 1993-1-2, cl. 4.2.5.1; Eq. (4.25) 32.6 min 27.1 min M. Hor´ aˇ cek et al. Thin-Walled Structures 192 (2023) 111117 16 6. Conclusions This paper described experimental fire tests on load-bearing thinwalled structural floor systems with unprotected cold-formed steel members. For these floor systems, found in storage halls, achieving the required 30-min fire resistance rating without fire protection presents an interest as they can be designed for a very low load ratio in the fire situation. However, the absence of experimental data and applicable calculation methods for such assemblies exposed to very high temperatures impedes their fire evaluation. Methods for the calculation of the fire resistance of cold-formed thin-walled structures in Europe [4,33] have seen limited benchmarking against fire tests on unprotected assemblies and these methods limit their range of validity to critical temperatures below those reached in an unprotected member after 30 min of fire exposure. In this context, the main findings from the experimental fire tests are presented here below. Temperatures measurements show that, after an initial lag of about 10 min, the temperature of the bottom flange of the unprotected beams closely follows the fire temperature, reaching 800 ◦C after 30 min. The Eurocode’s simple lumped mass thermal calculation method accurately captures the bottom flange temperature. The top flange, in contact with the chipboard, is cooler by about 150–200 ◦C at 30 min of exposure. The deflections of the floor systems remained limited during the first 20 min, then suddenly increased downwards. At 30 min, mid-span deflections were of the order of 100 mm (L/50) and 150 mm (L/33) in specimens A and B, respectively. These deflections are under the test standard threshold. However, the deflection rates in both specimens exceeded the threshold. The specimens would thus not qualify for a 30Table 7 Evaluation of deflection criterion for outer single primary beams (test configuration B). Fig. 25. Test specimen of configuration B from the unheated site after being exposed to the effects of fire for 30 min. M. Hor´ aˇ cek et al. Thin-Walled Structures 192 (2023) 111117 17 min fire resistance rating according to the fire resistance test standard [24]. These specimens were designed for 30-min fire resistance based on the degree of utilization and critical temperature, according to the Czech National Annex of EN1993-1-2 covering Class 4 cross-sections. Yet, this result confirms that the critical temperature approach is not intended to cover deformation criteria, and thereby, could not predict the exceedance of deflection rate thresholds in test standards. This has practical significance as it shows that the simple methods in the standard, while they could predict the thermal behavior, could not be relied on to predict the outcome of the standard qualification testing, because of differing failure criteria in terms of structural response. This effect is particularly pronounced for the studied assemblies, because given that they are unprotected and subject to minimal loads, their loadbearing capacity may be maintained at very high temperatures but large Fig. 26. Test specimen of configuration B from the heated side after burning and extinction. Fig. 27. Comparison of the steel temperature of primary beams in test configuration A. Fig. 28. Comparison of the steel temperature of primary beams in test configuration B. M. Hor´ aˇ cek et al. Thin-Walled Structures 192 (2023) 111117 18 thermally induced deformations develop. This may hinder the use of unprotected cold-formed steel assemblies. Given the high ductility of steel, it could be worth exploring whether the deflection rate threshold alone should signify the failure of the test or whether it should be considered in conjunction with the deflection threshold [32]. In fact, this research illustrates that test failure criteria developed for a range of applications could prevent the use of assemblies that were not considered at that time; as a matter of fact, the application of limiting temperature criteria (as suggested in the Czech National Annex for the simple calculation method, or applied in ASTM E119 testing in the United States) would invariably disqualify the unprotected cold-formed steel assemblies, although the full-scale experiments presented herein showed that load bearing capacity and deformation limits were still satisfied at close to 800 ◦C. CRediT authorship contribution statement Martin Hor´ aˇ cek: Writing – review & editing, Writing – original draft, Validation, Supervision, Project administration, Investigation, Formal analysis, Conceptualization, Methodology,Visualization. Thomas Gernay: Writing – review & editing, Writing – original draft, Supervision, Investigation, Conceptualization,Methodology,Formal analysis. Marcela Karmazínov´ a: Writing – review & editing, Writing – original draft, Supervision, Conceptualization. Marek Hladík: Supervision, Project administration, Funding acquisition,Conceptualization. Zbynˇ ek P¨ offel: Writing – original draft, Validation, Methodology, Investigation, Formal analysis,Visualization. Ondˇ rej Peˇ sek: Writing – original draft, Visualization, Investigation. Ivan Bal´ azs: Writing – original draft, Visualization. Declaration of competing interest The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Martn Hor´ aˇ cek reports financial support was provided by Technology Agency of the Czech Republic. Marcela Karmazínov´ a reports financial support was provided by Technology Agency of the Czech Republic. Marek Hladík reports financial support was provided by Technology Agency of the Czech Republic. Zbynˇ ek P¨ offel reports financial support was provided by Technology Agency of the Czech Republic. Ondˇ rej Peˇ sek reports financial support was provided by Technology Agency of Fig. 29. Comparison of the steel temperature of secondary beams in test configuration A. Fig. 30. Comparison of the steel temperature of secondary beams in test configuration B. Table 8 Overview of evaluated deflection and deflection rate criteria for primary beams. Fire test configuration Primary beam Time when the limit value was reached for: (a) deflection Eq. (1) (b) deflection rate Eq. (2) (c) deflection rate (only for d≥D/2) A Internal Not reached 26 min 29 min Outer Not reached 24 min 28 min B Outer Not reached 22 min 26 min M. Hor´ aˇ cek et al. Thin-Walled Structures 192 (2023) 111117 19 the Czech Republic. Ivan Bal´ azs reports financial support was provided by Technology Agency of the Czech Republic. Data availability Data will be made available on request Acknowledgments The authors gratefully acknowledge the Technology Agency of the Czech Republic (www.tacr.cz) for its support under the framework of research project FW01010206. The authors also gratefully acknowledge Professor Benjamin W. Schafer for arranging the contact between Dr. Martin Hor´ aˇ cek from the Institute of Metal and Timber Structures at the Brno University of Technology and Dr. Thomas Gernay from the Department of Civil and System Engineering at Johns Hopkins University, which led to successful cooperation in the preparation of this article. References [1] B. Schafer, Cold-formed steel structures around the world, Steel Constr. 4 (2011) 141–149, https://doi.org/10.1002/stco.201110019. [2] T. Ranawaka, M. Mahendran, Experimental study of the mechanical properties of light gauge cold-formed steels at elevated temperatures, Fire Saf. J. 44 (2009) 219–229, https://doi.org/10.1016/j.firesaf.2008.06.006. [3] Y. Yu, L. Lan, F. Ding, L. Wang, Mechanical properties of hot-rolled and coldformed steels after exposure to elevated temperature: A review, Constr. Build. Mater. 213 (2019) 360–376, https://doi.org/10.1016/j.conbuildmat.2019.04.062. [4] EN 1993 −1−2 eurocode 3: design of steel structures: part 1-2: general rules – structural fire design, 2005. [5] X. Yan, J. Abreu, R. Glauz, B. Schafer, T. Gernay, Simple three-coefficient equation for temperature-dependent mechanical properties of cold-formed steels, J. Struct. Eng. 147 (2021), https://doi.org/10.1061/(ASCE)ST.1943-541X.0002990. [6] F. Arrais, N. Lopes, P. Vila Real, Behaviour and resistance of cold-formed steel beams with lipped channel sections under fire conditions, J. Struct. Fire Eng. 7 (2016) 365–387, https://doi.org/10.1108/JSFE-12-2016-025. [7] L. Laím, J. Rodrigues, Fire design methodologies for cold-formed steel beams made with open and closed cross-sections, Eng. Struct. 171 (2018) 759–778, https://doi. org/10.1016/j.engstruct.2018.06.030. [8] S. Gunalan, P. Kolarkar, M. Mahendran, Experimental study of load bearing coldformed steel wall systems under fire conditions, Thin-Walled Struct. 65 (2013) 72–92, https://doi.org/10.1016/j.tws.2013.01.005. [9] S. Kesawan, M. Mahendran, Fire tests of load-bearing LSF walls made of hollow flange channel sections, J. Constr. Steel Res. 115 (2015) 191–205, https://doi.org/ 10.1016/j.jcsr.2015.07.020. [10] Y. Tao, M. Mahendran, A. Ariyanayagam, Fire tests of cold-formed steel walls made of hollow section studs, J. Construct. Steel Res. 178 (2021), https://doi.org/ 10.1016/j.jcsr.2020.106495. [11] Y. Zhong, A. Su, O. Zhao, Post-fire local buckling behaviour of cold-formed S700 high strength steel circular hollow sections under axial compression: Experiments, modelling and design, Thin-Walled Struct. 184 (2023), https://doi.org/10.1016/j. tws.2022.110511. [12] J. Yang, W. Wang, Y. Shi, L. Xu, Experimental study on fire resistance of coldformed steel built-up box columns, Thin-Walled Struct. 147 (2020), https://doi. org/10.1016/j.tws.2019.106564. [13] J. Yang, X. Zhou, W. Wang, L. Xu, Y. Shi, Fire resistance of box-shape cold-formed steel built-up columns failing in global buckling: Test, simulation and design, ThinWalled Struct. 183 (2023), https://doi.org/10.1016/j.tws.2022.110433. [14] L. Laím, J. Rodrigues, H. Craveiro, Flexural behaviour of beams made of coldformed steel sigma-shaped sections at ambient and fire conditions, Thin-Walled Struct. 87 (2015) 53–65, https://doi.org/10.1016/j.tws.2014.11.004. [15] L. Laím, J. Rodrigues, L. da Silva, Experimental analysis on cold-formed steel beams subjected to fire, Thin-Walled Struct. 74 (2014) 104–117, https://doi.org/ 10.1016/j.tws.2013.09.006. [16] M. Pandey, B. Young, RHS-to-RHS cold-formed S960 steel fire exposed X-joints: Structural behaviour and design, Thin-Walled Struct. 188 (2023), https://doi.org/ 10.1016/j.tws.2023.110711. [17] M. Pandey, B. Young, Design of cold-formed high strength steel CHS-to-RHS Tand X-joints at elevated temperatures, Thin-Walled Struct. 189 (2023), https://doi.org/ 10.1016/j.tws.2023.110728. [18] M. Pandey, B. Young, Post-fire mechanical response of high strength steels, ThinWalled Struct. 164 (2021), https://doi.org/10.1016/j.tws.2021.107606. [19] B. Baleshan, M. Mahendran, Numerical study of high strength LSF floor systems in fire, Thin-Walled Struct. 101 (2016) 85–99, https://doi.org/10.1016/j. tws.2015.12.018. [20] B. Baleshan, M. Mahendran, Experimental study of light gauge steel framing floor systems under fire conditions, Adv. Struct. Eng. 20 (2017) 426–445, https://doi. org/10.1177/1369433216653508. [21] E. Steau, M. Mahendran, Fire resistance behaviour of LSF floor-ceiling configurations, Thin-Walled Struct. 156 (2020), https://doi.org/10.1016/j. tws.2020.106860. [22] T. Abeysiriwardena, M. Peiris, M. Mahendran, Behaviour of stud-to-sheathing fastener connections in LSF walls at elevated temperatures, Eng. Struct. 238 (2021), https://doi.org/10.1016/j.engstruct.2021.112224. [23] A. Hamerlinck, Steel Design 2 - Fire: Fire Safety and Fire Resistant Design of Steel Structures for Buildings According to Eurocode, Vol. 3, 1st edition, Bouwen met Staal, Zoetermeer, 2021. [24] EN 1363-1 fire resistance tests – Part 1: general requirements, 2020. [25] EN 1365-2 fire resistance tests for loadbearing elements – Part 2: floors and roofs, 2014. [26] EN 1993 −1−3 eurocode 3: design of steel structures: Part 1-3: general rules – supplementary rules for cold-formed members and sheeting, 2006. [27] EN 1993-1-5: eurocode 3: design of steel structures - Part 1-5: general rules - plated structural elements, 2006. [28] M. Naser, N. Degtyareva, Temperature-induced instability in cold-formed steel beams with slotted webs subject to shear, Thin-Walled Struct. 136 (2019) 333–352, https://doi.org/10.1016/j.tws.2018.12.030. [29] EN 1993 −1−1 eurocode 3: design of steel structures: Part 1-1: general rules and rules for buildings, 2005. [30] EN 1990 Eurocode – Basis of Structural Design, European Committee for Standardization, 2002. [31] EN 1991 −1−2 eurocode 1: action on structures: Part 1-2: general actions – actions on structures exposed to fire, 2002. [32] F. Dumont, E. Wellens, T. Gernay, J. Franssen, Loadbearing capacity criteria in fire resistance testing, Mater. Struct. 49 (2016) 4565–4581, https://doi.org/10.1617/ s11527-016-0807-7. [33] J. Franssen, P. Vila Real, Fire Design of Steel Structures: Eurocode 1: Actions on Structures, Part 1-2: General Actions - Actions on Structures Exposed to Fire ; Eurocode 3: Design of Steel Structures, Part 1-2: General Rules - Structural Fire Design, 2nd edition, Ernst, Berlin, 2015. M. Hor´ aˇ cek et al.