J. Cent. South Univ. (2012) 19: 2054-2060
DOI: 10.1007/s11771-012-1244-6
Structural behaviors of steel roof truss exposed to pool fire
CHEN Chang-kun(陈长坤)1, 2, ZHANG Wei(张威)1
1. Institute of Disaster Prevention Science and Safety Technology,
Central South University, Changsha 410075, China;
2. State Key Laboratory of Fire Science (University of Science and Technology of China), Hefei 230026, China
? Central South University Press and Springer-Verlag Berlin Heidelberg 2012
Abstract: Experimental research was conducted to study the structural behaviors of a steel roof truss model without fire-proof coating under pool fire conditions. The data of temperature distribution and displacements of typical members were obtained. It is found that the temperature distribution of environment inside the structure, which is found to be in accordance with the multi-zone model with height, has a decisive effect on the temperature evolution of steel members. Besides, it can also be observed that due to the restriction and coordination among the truss members in the localized fire, the maximum relative deflection, which occurs at the mid-span of the top chord, is relatively slight and has not exceeded 1 mm under experimental conditions. On the other hand, the column experiences a notable thermal expansion during the test. Then, a finite element model is presented and validated by the test results.
Key words: steel roof truss; fire test; multi-zone model; finite element analysis
1 Introduction
Steel roof truss is a common type of steel structure, which is widely used in modern buildings [1]. At present, many studies have mainly concentrated on construction scheme, fatigue analysis, seismic behavior and dynamic response of steel roof truss at ambient temperature. For example, NUNO and DINAR [2] presented a study on elastic buckling and second-order behavior of pitched-roof steel frames. The proposed concepts were illustrated through the presentation of numerical results. ATTILA and MIKL?S [3] conducted a series of experiments to investigate the stability and ductility behavior of a space-truss roof system. CAGLAYAN and YUKSEL [4] made an investigation of the collapse of a Mero space truss roof structure in Turkey. SAKA [5] presented an optimum design of pitched roof steel frames by genetic algorithm. ASGARIAN and MORADI [6] discussed seismic response of steel braced frames with shape memory alloy braces. These studies have played a positive role in the promotion and application of steel roof truss as the scientific understanding of the mechanical performance of the structure at ambient temperature was deepened through these researches.
However, mechanical properties of steel deteriorate at elevated temperatures observably and for conventional steel the yield strength at 700 °C is less than 23% of the specified value at room temperature [7]. Thus, it is necessary to study the response of steel structure under elevated temperatures. A notable case concerned is the Cardington fire test, which was performed in order to investigate the global structural behavior of steel-framed building under realistic fire conditions [8]. In particular, structural responses of steel constructions in fire were more noticeable since the World Trade Center (WTC) Twin Towers collapsed in 2001 [9]. Parametric analysis and experimental research were performed to study the mechanical behavior of steel planar tubular trusses under elevated temperatures [10-11]. ZHAO and SHEN [12] conducted a total of three tests on steel frames under different load levels and different heating processes in order to study the structural fire behavior of steel frame structure. However, most of these earlier studies were mainly carried out in a fire test furnace, in which the temperature rise might be different from a nature fire. Actually, the mechanical response behavior of steel roof truss exposed to nature fire might be different from that in a fire test furnace [13-14]. In addition, the structural behavior of steel roof truss might be more complicated since the steel roof truss is composed of more members and joints compared to the ordinary steel frame structure. Thus, experimental and computational research is conducted in this work to investigate the mechanical behavior and temperature distribution of steel roof truss exposed to pool fire, which is considered to enable development of scientific understanding and rational design methods for the structure in fire.
2 Test set-up
2.1 Description of test model
Figure 1 shows the test model, which is a rectangle steel roof truss with plan dimension of 0.8 m×1.5 m. It is composed of three transverse frames with spacing of 0.4 m. The height of the model is 1.8 m. Two types of frames named Frame A and Frame B (including Frame B1 and Frame B2) are used in the test model. Frame B1 and Frame B2 were placed on either side of Frame A. Frame B has a span of 1.5 m with five panels in each truss, and the width of each panel is 0.3 m, as illustrated in Fig. 2. The floor slab is 2 mm steel sheet stiffened by steel bars with spacing of 0.3 m. The details of model members are shown in Table 1. The model is bolted onto the concrete foundation (see Fig. 3). Two adjacent frames are interconnected with longitudinal beams. Web members and chords in a single truss are connected by fully welded joints, which can be considered as rigid connection. Gusset plates and high strength bolts are adopted to connect chords with external columns, as shown in Fig. 4.
Fig. 1 General view of experimental model
All the members were fabricated from Grade Q235B which is commonly used in Chinese construction industry. In order to prepare for the numerical analysis of steel roof truss, steel tensile coupon tests were carried out to determine the mechanical properties of the steel members at ambient temperature. The tested elastic modulus and Poisson ratio are 2.0×105 N/m2 and 0.3, respectively, and the measured tensile strength is 330 N/mm2.
Fig. 2 Schematic arrangement of test (Unit: mm)
Table 1 Details of members used in experimental model
Fig. 3 Photographs of connection between test model and foundation: (a) Connection between Frame A and foundation; (b) Connection between Frame B and foundation
Fig. 4 Photographs of joints: (a) Connection between chord and external column; (b) Connection between web member and chord
2.2 Test conditions and procedure
Pool fire is used as the fire source to simulate real fire conditions during the test. The length, width and height of the fuel pan are 0.3, 0.3 and 0.05 m, respectively. The fuel pan filled with 2.6 kg of methanol was placed at the center of the ground during the test. During the test, members of the test model were unprotected. It should be noted that the test model was enclosed by calcium silicate board for the purpose of getting an appropriate thermal environment. Ventilation opening with dimensions of 0.4 m (width)×0.6 m (height) was reserved at Frame B1, as shown in Fig. 5.
Fig. 5 Photo of steel roof truss during test
Before the test model was exposed to the fire, 200 kg of weights were applied manually to the floor slab to simulate uniform distribution load condition. That is to say, the imposed load was 1.633 kN/m2. The loading was then maintained at the same level throughout the fire test. Then, the fire was ignited, and the data of temperature distribution and displacements were recorded. The test was terminated at about 6 min after the fire went out.
2.3 Test instrumentation
The test members were instrumented with thermocouples and displacement transducers to monitor their thermal and mechanical response during the fire test. Specifically, seven displacement transducers were installed on the top chord of Frame B1 and the top of the column to monitor the vertical deflections.
A total of twenty-three thermocouples were installed on typical members to record detailed temperature distributions during the test. Particularly, seven thermocouples were arranged on typical members of Frame B1, while ten other thermocouples were fixed at typical positions of Frame B2. In addition, six more thermocouples were attached to the column of Frame A. The distance between two adjacent thermocouples of Frame A was 0.3 m.
All thermocouples and displacement transducers were connected to the self-developed acquisition system and the data were recorded at 3 s interval. The locations of the different types of measuring devices are shown in Fig. 6. Besides, several cameras were used to record the phenomenon throughout the test.
3 Experimental results
Data obtained from the fire test are used to study the temperature distribution and the overall performance of steel roof truss under pool fire conditions.
Figure 7(a) shows the temperature evolution of the column of Frame A. As mentioned previously, the distance between two adjacent thermocouples is 0.3 m. It is found that the temperature of steel rises along with the increase of the height. The temperature of thermocouple 10 (T10) located at the topmost is the highest, attaining 270 °C at about 1 100 s. However, the temperature of T13, which is located at the bottom of the column, keeps almost unchanged throughout the test. This is because the temperature distribution of the space inside the structure accords with the multi-zone model, which has been found by KEISHI et al [15]. In particular, the space in the structure can be divided into hot smoke layer at the top and cool air layer at the bottom. With the descending of the smoke layer, the temperature of the environment drops at the lower position. It is important to note that the thermal environment inside the structure has a decisive effect on the temperature evolution of steel members. Thus, the temperature of steel members also complies with the multi-zone model, which is well reflected during the test.
Temperature evolution of typical members of Frame B2 is shown in Fig. 7(b). The trends of temperature evolution of typical members are similar. The temperature recorded during the test rises first and then decreases since the fire source begins to abate at about 1 100 s. Among these measuring points, the temperature of T5 is the highest, about 320 °C. The temperatures of steel members recorded by thermocouples at the same altitude are almost the same. It can be inferred that the temperature distribution of steel members of Frame B2 also accords with the multi-zone model. It is important to note that the temperatures of top chord (i.e. T3, T6 and T16) are slightly lower than the temperatures of web members and bottom chord though the altitude of top chord is higher. The possible reason might be that heat of the hot smoke is exchanged with the cooler outside ambient environment through the steel floor sheet.
The temperatures of typical members of Frame B1 recorded during the fire test are plotted in Fig. 7(c). The data of T19, T20 and T21 are not adopted for these thermocouples slip out during the test. The temperatures of T17 and T18, T22 and T23 are approximately the same. Also, the temperatures of T17 and T18 are a little higher than the temperatures of T22 and T23. As mentioned in Figs. 7(a) and (b), it can be concluded that the temperature evolution of steel members is highly determined by the temperature distribution of thermal environment in the structure, which is found to be in accordance with the multi-zone model with height.
Fig. 6 Locations of measurement devices (Unit: mm)
Fig. 7 Temperature evolutions of typical positions: (a) Column of Frame A; (b) Frame B2; (c) Frame B1
Figure 8 shows the vertical displacements of representative positions of Frame B1. It should be pointed out that the upward vertical displacement is designated to be positive and the relative deflection is the measured displacement subtracting the displacement of the column. The data recorded during the test are demonstrated partly due to the symmetry of the structure. The displacement of D1 located at the top of the column experiences an increase and then reduces throughout the test. The key factor contributing to the displacement of the column is thermal expansion. In particular, the maximum relative deflection induced by elevated temperatures still occurs at the mid-span section (see D4 in Fig.8). However, as illustrated in Fig. 8, the displacements of D3 and D4 are basically similar due to the braced effect of web members. It could be inferred that the relative deflection of the top chord decreases as a result of the coordination of steel members of steel roof truss, which indicates that the truss members have a significant influence on the performance of steel roof truss under elevated temperatures.
Fig. 8 Vertical displacement of representative positions of Frame B1
4 Finite element analysis
In order to present a reasonable finite element model for the further study in the parameter analysis of steel roof truss under fire conditions, a numerical simulation on the test model was carried out using the finite element analysis program ANSYS. Regarding the finite element model, a finite strain beam (Beam 188) was used for the discretization of both columns and other steel members of the test model. Four-node finite strain shell (Shell 181) was adopted to simulate the steel sheet. The dimensions of the finite element model were kept consistent with the test model [16]. For the convenient purpose of calculation, the residual stresses resulted from the welding were neglected since their effect on the static strength of the truss has been found to be insignificant according to the findings of LEE and LLEWELYN-PARRY [17]. The fixed connection was used between the finite element model and the foundation. Besides, the connections between top chords and columns were considered to be hinged joints. Rigid connection was utilized for the simulation of connections between chords and web members. In the calculation, steel material properties at elevated temperature were in agreement with Eurocode 3 [18]. Particularly, the value of expansion coefficient of steel was designated as 1.2×10-5 [19]. The distribution of temperature was constructed using the temperature data recorded during the experiment. The loading and boundary conditions for the finite element model of the steel roof truss were the same as those of the test model. The finite element model is shown in Fig. 9.
The results of the relative deflection of D3 and D4 (see Fig. 8) were compared between the numerical analysis and the experimental results, as illustrated in Fig. 10. As can be seen from Fig. 10, acceptable agreement is obtained between the finite element model curves predicted by ANSYS and experimental results. However, a certain difference exists between numerical prediction and test results, which might result for four reasons as follows: The first is that the boundary conditions and constraints utilized in the simulation were in idealization compared to the experiment; The second is that, the temperatures used in the simulation were simplified at a certain degree for only representative temperatures of steel members were recorded due to the limitations in resource and time during the test; The third is that the loading condition was impossible to achieve a perfect uniform distribution during the test compared to the numerical study; The fourth is that the actual material properties under elevated temperatures in the test had some differences from Eurocode 3.
Fig. 9 Finite element model of steel roof truss
Fig. 10 Comparison between test results and numerical study: (a) Relative deflection of D4; (b) Relative deflection of D3
It could be concluded that the finite element model presented in this work is acceptable and has considerable ability of predicting structural behaviors of steel roof truss in fire. Specially, the model could be used for the future parameter study to predict the performance of steel roof truss during the fire exposure.
5 Conclusions
1) The temperature of environment in the structure has a decisive effect on the temperature evolution of steel members. Under pool fire conditions, the temperature distribution of the environment in the structure accords with the multi-zone model with height. Thus, the temperature of steel members accords with the multi-zone model accordingly.
2) The experiment is considered to be successful in giving a clear picture of temperature evolution and mechanical response of the structure though the failure mode is not obtained during the test. External columns experience a remarkable thermal expansion during the test, and the maximum relative deflection occurs at the mid-span cross section of the top chord. In particular, the truss members have a significant influence on the performance of steel roof truss in fire.
3) The comparison between the test results and the computational study shows acceptable agreement. Based on this, the result of finite element model with beam and shell elements presented has a relatively considerable precision. Thus, the model might be used in the further study on parameter analysis of steel roof truss under fire conditions.
4) It is worthy to note that the test is mainly concentrated on the structural behavior subjected to a relatively small-scale localized fire. The case of high-power fire should also be explored further in the future for a comprehensive understanding on the fire performance of the structure.
References
[1] WANG Xiu-li. Analysis and concept design of large span spatial structure [M]. Beijing: China Machine Press, 2008: 223-236. (in Chinese)
[2] NUNO S, DINAR C. Elastic buckling and second-order behavior of pitched-roof steel frames [J]. Journal of Constructional Steel Research, 2007, 63(6): 804-818.
[3] ATTILA F, MILL?S I. Experimentally analyzed stability and ductility behavior of a space-truss roof system [J]. Thin-Walled Structures, 2004, 42(2): 309-320.
[4] CAGLAYAN O, YUKSEL E. Experimental and finite element investigations on the collapse of a Mero space truss roof structure— A case study [J]. Engineering Failure Analysis, 2008, 15(5): 458-470.
[5] SAKA M P. Optimum design of pitched roof steel frames with haunched rafters by genetic algorithm [J]. Computers and Structures, 2003, 81(18/19): 1967-1978.
[6] ASGARIAN B, MORADI S. Seismic response of steel braced frames with shape memory alloy braces [J]. Journal of Constructional Steel Research, 2011, 67(1): 65-74.
[7] WANG Y C. Steel and composite structures, behavior and design for fire safety [M]. London: Spon Press, 2002: 106-113.
[8] WANG Y C. An analysis of the global structural behavior of the Cardington steel-framed building during the two BRE fire tests [J]. Engineering Structures, 2000, 22(5): 401-412.
[9] USMANI A S, CHUNG Y C, TORERO J L. How did the WTC towers collapse: A new theory [J]. Fire Safety Journal, 2003, 38(6): 501-533.
[10] JIN Meng, ZHAO Jin-cheng, LIU Ming-lu, CHANG Jing. Parametric analysis of mechanical behavior of steel planar tubular truss under fire [J]. Journal of Constructional Steel Research, 2011, 67(1): 75-83.
[11] LIU Ming-lu, ZHAO Jin-cheng, JIN Meng. An experimental study of the mechanical behavior of steel planar tubular trusses in a fire [J]. Journal of Constructional Steel Research, 2010, 66(4): 504-511.
[12] ZHAO Jin-cheng, SHEN Zu-yan. Experimental studies of the behavior of unprotected steel frames in fire [J]. Journal of Constructional Steel Research, 1999, 50(2): 137-150.
[13] CHEN Chang-kun, ZHANG Wei. Experimental study of the mechanical behavior of steel staggered truss system under pool fire conditions [J]. Thin-Walled Structures, 2011, 49(11): 1442-1451.
[14] CHEN Chang-kun, ZHANG Dong, ZHANG Wei, SHEN Bing-yin. Experimental behaviors of steel staggered-truss system exposed to fire under lateral force [J]. International Journal of Steel Structures, 2012, 12(1): 59-70.
[15] KEISHI S, KAZUNORI H, TAKEYOSHI T. A multi-layer zone model for predicting fire behavior in a single room [C]// Proceedings of the Seventh International Symposium. Gaithersburg: Fire Safety Science, 2002: 851-862.
[16] YANG X L, YIN J H. Slope stability analysis with nonlinear failure criterion [J]. Journal of Engineering Mechanics, 2004, 130(3): 267-273.
[17] LEE M M K, LLEWELYN-Parry A. Strength of ring stiffened tubular T-joints in offshore structures: A numerical parametric study [J]. Journal of Constructional Steel Research, 1999, 51(3): 239-264.
[18] Eurocode 3: EN1993-1-2: Design of steel structures-Part 1-2: General rules-Structural fire design [S]. Brussels (Belgium): Commission of European Communities, 2005.
[19] LI Guo-qiang, HAN Lin-hai, LOU Guo-biao, JIANG Shou-chao. Fire resistant design of steel structure and steel-concrete composite structure [J]. Beijing: China Architectureand Building Press, 2006: 89-91. (in Chinese).
(Edited by YANG Bing)
Foundation item: Project(50706059) supported by the National Natural Science Foundation of China; Project(HZ2009-KF05) supported by Open Fund of State Key Laboratory of Fire Science of University of Science and Technology in China; Project supported by the Fundamental Research Funds for the Central Universities of China
Received date: 2011-07-26; Accepted date: 2011-11-14
Corresponding author: CHEN Chang-kun, Professor, PhD; Tel: +86-731-82656625; E-mail: cckchen@csu.edu.cn
