Temperature of a partially embedded connection subjected to fire Jiˇ rı ´ Chlouba, Frantiˇ sek Wald n Czech Technical University in Prague, Czech Republic article info Article history: Received 29 October 2011 Received in revised form 7 August 2012 Accepted 13 August 2012 Available online 8 September 2012 Keywords: Fire design Composite steel and concrete structure Heat transfer Connection design Partially embedded connection abstract When subject to fire, structural steel and connectors lose both their strength and stiffness. Structures expand when heated and contract on cooling. Furthermore, the effect of restrained thermal movement can introduce high strains in both the steel member and the associated connections. Fire tests on steel structures have shown that the temperature within the connections is lower compared to the connected steel members. The beneficial effect of the partial embedding of a connection in a concrete slab for its resistance in fire conditions was already intuitively utilised by structural engineers. Three sets of tests provide the measuring of the accuracy of the developed numerical and analytical model for heat transfer and temperature distribution in the partially embedded connections. The connection shadow factor is proposed further to allow the prediction of temperatures in connections from the calculated gas temperature during a post-flashover fire. & 2012 Elsevier Ltd. All rights reserved. 1. Introduction The fire resistance of structures has been traditionally verified by experiments on fire protected or unprotected elements. Currently, the test results are substituted for a member analysis of structural elements. Higher economy can be achieved by verifying the overall structural behaviour; particularly when accounting for the membrane action of the composite steel and concrete floors of multi-storey buildings. In such cases, structures may meet the requirements also being only partially fire pro- tected, see [1]. There are two approaches to the fire design of steel connec- tions. Using the first approach, fire protection is applied to the member and its connections. The level of fire protection is based on the protection of the connected members. In the design, the degree of efficiency is considered, which may vary between the connections and connected members. More detailed second approach is uses an application of the component procedure described in EN1993-1-8, see [2], together with a method for calculating the behaviour of welds and bolts at elevated tem- peratures. Using this approach, the connection moment, shear and axial capacity can be evaluated at elevated temperatures with a good level of accuracy by simple or advanced models of mechanical behaviour, see [3]. The accuracy of the design depends on prediction of the temperature of elements and of forces acting in the connection during the fire. The thermal conductivity of steel is high. Nevertheless, because of the concentration of material within the joint area, differential temperature distribution should be considered within the joint. Various temperature distributions have been proposed or used in experimental tests by several authors, see [4]. In an attempt to quantify the temperature distribution within a joint, tests have been done on several joint typologies, see [5]. One example for larger beams, a web temperature similar to the bottom flange temperature, is provided while for smaller beams, a smaller web temperature is provided, see [6]. Additionally, the presence of the concrete slab above the joint causes a reduction in beam top flange temperatures, see [7]. A detailed description can be found in the literature [8,9]. The values proposed in EN1993-1-2, see [10], are in agreement with these experimental results. However, these values are based on a nominal standard fire curve, if the fire follows other curves it is necessary to analyse the particular case using a numerical or experimental study, see [11,12]. According to EN1993-1-2, the temperature of a joint may be assessed using the local massivity value A/V of the joint compo- nents. As a simplification, a uniform distributed temperature may be assumed within the joint; this temperature may be calculated using the maximum value of the ratios A/V of the adjacent steel members. For beam-to-column and beam-to-beam joints, where the beams support any type of concrete floor, the temperature may be obtained from the temperature of the bottom flange at midspan, see Fig. 1. The temperature of the joint components at its height h k may be determined, see Annex D of EN1993-1-2, as follows. The depth of the beam h b is less or equal to 400 mm y k ¼ 0:88 y 0 ½10:3ðh k =h b Þ ð1Þ Contents lists available at SciVerse ScienceDirect journal homepage: www.elsevier.com/locate/firesaf Fire Safety Journal 0379-7112/$ - see front matter & 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.firesaf.2012.08.008 n Corresponding author. Tel.: þ420 224354757; fax: þ420 233334766. E-mail addresses: jiri.chlouba@fsv.cvut.cz (J. Chlouba), wald@fsv.cvut.cz (F. Wald). Fire Safety Journal 54 (2012) 121–129