Elsevier

Thin-Walled Structures

Volume 98, Part A, January 2016, Pages 143-153
Thin-Walled Structures

Post-earthquake fire behavior of welded steel I-beam to hollow column connections: An experimental investigation

https://doi.org/10.1016/j.tws.2015.03.032Get rights and content

Highlights

  • Tests on unstiffened welded steel connections subjected to post-earthquake fire.

  • The higher the pre-damage level is, the more reduction in connection strength is.

  • Heat transfer analysis revealed temperature distribution of the connection.

Abstract

Extreme events such as earthquake and fire can cause severe damage to building structures. The possible coupling of various extremes such as earthquake followed by fire is more destructive. It is necessary to capture the behavior of building structures subjected to the post-earthquake fire (PEF) for safety reasons. This study aims to investigate the behavior of unstiffened welded steel connections subjected to the PEF scenario. Experimental tests were carried out on two groups of steel I-beam to hollow column connections with different column wall thicknesses under ISO fire including 4 connections with the same specifications within each group in which 3 connections were tested under the coupling of cyclic loading and subsequent fire while 1 connection was only subjected to fire. All undamaged and pre-damaged connections were tested in a gas furnace under the given constant static load which was about 30% of the ultimate monotonic loading strength of the connections. The furnace temperature, temperature distribution and the deflection of the beam were measured during the fire tests. It was found that the load-carrying capacity of the welded connections decreases significantly with the increase of the pre-damage level. Preliminary finite element modeling was carried out for a heat transfer analysis on the basis of uniform fire exposure as well, which showed nonuniform temperature distribution of the connections.

Introduction

Both earthquake and fire can cause severe damage. The ground motion due to earthquake often causes wide and huge damages to building structures. Usually, various degrees of damage from no damage to collapse can be observed on steel structures in earthquakes [1], [2], [3]. Especially in low to moderate seismic risk regions, such as Australia, the unprepared cities are also prone to damage by the low probability but high consequences earthquake events [4]. Damage of steel structures induced by earthquake could be observed as residual deformation, loss of fire protection material, fractures, etc. It was also shown that the pre-damage caused by cyclic loading or high-strain loading would affect the mechanical properties of mild steel at elevated temperature [5], [6], [7]. Experimental tests on the simple welded and double-angle bolted steel I-beam to tubular column connections under cyclic loading were reported by Song et al. [8], [9], [10], in which the damage was observed including not only the residual deformation but also physical damage such as local buckling and steel fractures. Various cummulative damages were also shown from cyclic loading tests of other types of steel and composite connections [11], [12]. The partially damaged structures, which survive after earthquake, are much vulnerable under secondary actions such as fire (PEF) [13]. In other words, with the pre-existing damage caused by seismic loading, fire resisting capacity of steel structures decreases.

Research on the post-earthquake fire behavior of structures is very limited. Zaharia and Pintea [14], [15] discussed the damage level induced by the earthquake on the fire resistance of unprotected steel moment resisting frames. Damage patterns of sprayed fire resistive material on steel structural components due to a strong seismic event were presented by Braxtan and Pessiki [16]. Della Corte et al. [17] investigated the fire resistance of the steel moment-resisting frames. It was found that the pre-damage caused by an earthquake is a key factor which influences the behavior of steel structures [18], [19] or steel-concrete composite joints [20], [21], [22], [23], [24], [25] subjected to PEF. It indicated that when the residual deformation is not large enough the connection behavior under fire would not be obviously affected by the induced pre-damage. Preliminary FE modeling investigation of steel connections exposed to post-earthquake fire was presented by Song et al. [26], in which reduction of the fire resistance was analyzed considering only the residual deformation and residual strain/stress states as earthquake induced pre-damage. Moreover, Taylor [27] examined the factors involved in determining the safety of tall buildings exposing to PEF. The material degradation and heat penetration of a concrete building structure after earthquake were investigated by Mostafaei and Kabeyasawa [28]. Collier [29] presented the project endeavored to quantify the resultant reduction in the fire resistance of a series of plasterboard lined lightweight timber and steel-framed walls.

Although the above mentioned investigations are related to the structure behavior under PEF, however the experimental investigations on the structural components are still limited since various structure components and several factors might be involved. In this study, experimental investigation is carried out on the thermal behavior of I-beam to hollow steel column connections affected by pre-existing damages induced by seismic loading. Simple welded steel I-beam to tubular column connections were fabricated for the proposed cyclic loading and fire tests. Temperature distribution over various components of the connections at elevated temperature were analyzed and compared with the heat transfer analysis results on the basis of uniform fire exposure. Finally, the fire resistance capacity of the undamaged and damaged connections were compared and analyzed. The experimental results reported in this study provide fundamental understanding of effect of the seismic loading induced damage on the load-carrying capacity of the welded steel connections. It indicates the necessity to evaluate structure behavior under the post-earthquake fire scenarios for safety reasons.

Section snippets

Specimens and fabrication

Two groups of welded connections were fabricated for the experimental investigation in this study. Each group contains 4 connections with the same configuration details. Details of the unstiffened welded connections are shown in Fig. 1. The cross-section of all I-beams are made of 200UB22.3 (Grade 250 steel) [30]. The square tubular columns, with the cross-section of 200 mm×200 mm made of the Grade 350 steel, were adopted. The tubular columns of the thickness of 5 mm and 9 mm are adopted which form

Pre-damage loading tests

According to the results of the cyclic loading tests [8], the pre-damage induced to the connections, is mainly attributed to the fracture and residual deformation, which increases cumulatively with the increase of the loading cycles. The load vs. displacement relationships of the pre-damage loading phase and the skeleton curve obtained from the cyclic loading tests [8] are shown in Fig. 4. It can be seen that the pre-damage loading process follows the cyclic loading skeleton curve well except

Heat transfer analysis and temperature distribution

By adopting the measured furnace temperature as input data, a heat transfer analysis considering uniform exposure was carried out using ABAQUS [33]. Thermal properties of the steel material were assumed not affected by seismic loading while the effect of the heat penetration due to steel fracture was not considered. The convection heat transfer coefficient was taken as 25 W/m2K while the radiation heat transfer coefficient was assumed to be 0.5 W/m2K [34]. The specific heat (cs) and conductivity

Conclusions

Experimental investigation of post-earthquake behavior of a series of unstiffened welded steel I-beam to tubular column connections under fire was presented in this paper. Two groups of connections with different column wall thicknesses were tested in a gas furnace set with ISO-834 standard fire curve. Although the temperature measured in the middle of the furnace was a little lower than the ISO-834 standard fire curve due to the sealing conditions, it provided relatively similar fire scenarios

Acknowledgment

The research work presented in this paper was supported by Australian Research Council through a Discovery Project (DP1096454) awarded to the second author. The authors would like to express gratitude to the laboratory manager Mr Long Goh who prepared materials and upgraded equipment for the experimental tests.

References (35)

  • J.L. Wilson et al.

    Development of the New Australian earthquake loading standard

    Electron J Struct Eng

    (2008)
  • Q.Y. Song et al.

    Performance of unstiffened welded Steel I-beam to hollow tubular column connections under seismic loading

    Int J Struct Stab Dyn

    (2015)
  • Q.Y. Song et al.

    Performance of double-angle bolted Steel I-Beam to hollow square column connections under static and cyclic loadings

    Int J Struct Stab Dyn

    (2015)
  • Q.Y. Song et al.

    Seismic behaviour of welded and bolted I-beam to box column connections

    (2013)
  • Chen S, Lee GC, Shinozuka M. Hazard mitigation for earthquake and subsequent fire. In: Proceedings of the 2004 ANCER...
  • R. Zaharia et al.

    Fire after earthquake analysis of steel moment resisting frames

    Int J Steel Struct

    (2009)
  • Zaharia R, Pintea D. Advanced method for fire analysis of structures after earthquake.In: Proceedings of the 11th wseas...
  • Cited by (34)

    • Experimental study on high-strength Q460 steel extended end-plate connections at elevated temperatures

      2023, Journal of Constructional Steel Research
      Citation Excerpt :

      As for extended end-plate connections, the influence of load ratio at the beam end, stiffening ribs of the column web, end-plate thickness, and bolt details were investigated experimentally at elevated temperatures [8–14]. Song [15] has conducted beam-to-column connections experiments considering the coupling effect of earthquake and fire, and the influence of different pre-damage levels on connections was revealed. Qian [16] investigated the influence of beam axial pressure on the moment resistance and deformation capacity of extended end-plate connections at the temperature of 400 °C, 550 °C and 700 °C.

    View all citing articles on Scopus
    View full text