In many design codes for timber structures (e.g. Eurocode 5 and SIA 265), the stiffness of a connection is given by empirical equations for a single dowel-type fastener per shear plane. The global stiffness of the connection is then given by multiplication with the number of dowels and shear planes. In the codes cited above, the empirical equations to estimate stiffness only depend on two parameters, namely the dowel diameter and the wood density. The main difference between different codes and stiffness of different types of fasteners is the choice of the exponent on these two parameters. Development and background of empirical equations for stiffness in different codes were recently reported in Jockwer and Jorissen (2018), who analysed about one thousand double shear timber-to-timber connection tests to evaluate the influence of further parameters on stiffness, such as number of fasteners in a row, number of rows of fasteners and dowel slenderness. From this huge database, they have been able to estimate another empirical stiffness equation, that includes dowel slenderness as an additional parameter. Effects of the latter have even been reported in Lemaître et al. (2018), by using a phenomenological numerical model instead ofexperiments. Sandhaas and van de Kuilen (2017) reported that using the slip modulus Kser, calculated according to Eurocode 5 for stiffness prediction of multiple fastener joints, considerably overestimates the experimentally observed stiffness and they proposed to introduce an effective number of dowels in their design, which was also recommended in Jockwer and Jorissen (2018).

This paper continues the work presented in Lemaître et al. (2018) on strength and stiffness estimations of single-fastener connections using a beam-on-foundation (BOF) modelling. In Lemaître et al (2018), the beam-on-foundation model calculations were compared to design equations of Eurocode 5, i.e. the load-carrying capacity and slip modulus. By these comparisons, the validity of the method for the design of single-fastener connections was highlighted. Moreover, effects that are not explicitly covered in the empirical design equations, namely the influence of the dowel slenderness and the nonlinear dowel diameter on the slip modulus were demonstrated.

In the present paper, the same comparisons are made for multiple dowelled connections. Moreover, load distribution between dowels in this type of connection, which was shown to be non-uniform by Blass (1995), is studied by means of the BOF model. Different approaches to estimate the load distribution have been proposed by Cramer (1968), Lantos (1969) and Wilkinson (1986).