Experimental and numerical analyses of rounded dovetail timber connections (RDC) under fire conditions

Highlights

• Rounded dovetail timber connections are tested under fire conditions.

• Thermal flux inside the connection.

• Eurocode 5 does not take into account the heating inside the connection.

Abstract

The use of rounded dovetail connections has gained popularity in timber floor and ceiling structures due to the advance in computerized numeric control machinery. However, European building regulations for fire safety in timber structures from Eurocode 5 standard does not provide a specific method to calculate the fire performance of this kind of connections. In this work, two experimental fire tests were made to evaluate the fire performance and load-bearing capacity of this timber connection. A numerical coupled thermo-mechanical simulation of the tests was developed using the general advanced calculation methods proposed in annex B of Eurocode 5. The experimental tests showed that this connection is not able to accomplish with a R30 fire resistance class, which is the minimum requirement for lightweight timber frame assemblies. A loss of material caused by charring under the tenon of the connection leads to the failure. The general methods proposed in annex B of Eurocode 5 does not take into account the heating inside the connection. However, the simulation results showed an underestimation of the charring rate in the connection. A new simulation considering the thermal flux inside the connection was developed and it shown good agreement with the experimental tests.

Introduction

The fire performance of a timber structure is largely influenced by the behaviour of its connections [1]. The size of the cross section of a timber element decreases gradually under fire conditions. Eventually, this loss leads to the collapse of the element. The time between the ignition and the collapse of the element supporting an external load is defined as the load-bearing capacity (R) of the element [2] (see Table1).

The European timber construction code (Eurocode 5, part 1.2 [3]) provides methodologies to design timber connections able to withstand fire conditions. However, timber-to-timber carpentry connections are not included in this code. These connections join timber elements by cutting and fitting them, without nails, screws or bolts. The forces are transmitted from one piece to another through mortises and tenons, or notches and pins. The axial forces are transmitted through compressions and tangential forces [4].

This kind of connections were very time consuming and expensive, and consequently rarely used in building industry. In recent years, the use of carpentry connections has enjoyed a comeback due to advanced software packages and techniques to design and manufacture them. These include Computer Aided Design or Computer Aided Manufacturing (CAD/CAM) and Computer Numeric Control (CNC) [5].

The rounded dovetail connection (RDC) is a carpentry connection, particularly used in roof and floor frames. It transmits loads from secondary structural elements, such as joists, to primary structural elements such as beams. The mechanical behaviour of RDCs at ambient temperature is well known, and the critical parameters of its design have been studied [6], [7], [8]. The results show that these connections not only transfer vertical shear forces but also carry load in tension and bending. These works also show that the angle and height of the dovetail flange affect the structural behaviour of RDCs significantly. Furthermore, a probabilistic method was proposed for the improvement of RDC design. Tannert [9] conducted a series of experiments on RDC to study different methods to increase the stiffness of these structural elements. The research shows that stiffness is improved by increasing the size of the tenon or by reinforcing the joints.

Connections with dowel-type fasteners are commonly used in timber structures, and their mechanical behaviour under a fire event have been studied experimentally since the late 1970s and early 1980s [10]. In the late 1990s, fire tests on three-member timber-to-timber and steel-to-timber (with an internal steel plate) connections, and exposed to fire on all sides were performed [11], [12], [13]. These results were included in EN 1995-1-2:2004. Fire tests on other connections typologies were carried out in the 2000s by [14] and in the 2010s by [15], [16], [17]. However, the research into the performance of carpentry-type timber connections under fire conditions is still limited. There is little work on this topic, and the one that has been made, is all about dovetail connections [18], [19], [20]. No research about any other kind of carpentry-type timber connection has been found in scientific literature. Racher et al. [18] studied dovetail connections under fire conditions. All but one of the tests exceed 15 min of fire exposure. Zhang et al. [19] conducted several experimental tests on straight-line dovetail joints under fire conditions. The experimental tests include protected and unprotected dovetail joints. In both cases, a gap between the tenon and the mortise was identified. The results show that the existing gap had a negligible effect on heat transfer. Furthermore, the fire-retardant coating improves the performance and significantly increases the fire protection.

These previous works do not specify the compliance with the Eurocode 5 [3] requirements for timber elements under fire conditions (R).

Experimental tests under fire conditions are extremely expensive and complicated. Numerical analysis can be used to study different geometrical joints and thermo-mechanical parameters. In this context, Regueira et al. [20] developed a finite element (FE) model to predict the mechanical performance of RDC under fire conditions. This numerical analysis, which is based on the software ANSYS, was validated using experimental results, leading to two conclusions. Firstly, the numerical results show that the temperature on the sides of the joint is influenced by its geometry. Secondly, the performance of the RDC does not meet the R30 performance criteria for fire resistance. Zhang et al. [21] developed a 3D FE model to simulate the thermo-mechanical behaviour of dovetail joints under fire conditions. The approach was validated using experimental data. The gap between the mortise and tenon is not considered in the numerical model and there is a discrepancy regarding numerical and experimental results at the beginning of tests. Despite this, there is good agreement with the experimental results.

In this research work, the load-bearing capacity of a timber beam connected to a joist using an RDC is experimentally studied under fire conditions. Then, numerical simulations using the finite element method (FEM) are developed. Firstly, two thermal models are developed using different thermal boundary conditions to study the self-protection performance of the connection. Then coupled thermal-structural models are studied and compared with the experimental results