The aim of the experimental study presented herein is the assessment and quantification of the behavior of individual dowels in multi-dowel connections loaded by a bending moment. For this purpose, doubleshear, steel-to-timber connections with nine steel dowels arranged in different patterns and with different dowel diameters were tested in 4-point bending. In order to achieve a ductile behavior with up to 7° relative rotation, the connections were partly reinforced with self-tapping screws. The reinforcement did not influence the global load-deformation behavior, neither for dowel diameters of 12 mm nor for 20 mm, as long as cracking was not decisive. The deformation of the individual dowels was studied by means of a non-contact deformation measurement system. Thus, the crushing deformation, i.e. the deformation at the steel plate, and the bending deformation of the dowels could be quantified. In case of 12 mm dowels, the bending deformation was larger than the crushing deformation, while it was smaller in case of 20 mm dowels. Moreover, dowels loaded parallel to the grain showed larger bending deformations than dowels loaded perpendicular to the grain. This indicates that the loading of the individual dowels in the connection differs, depending on their location.
In this paper, a non-linear procedure for the seismic design of metal connections in cross-laminated timber (CLT) walls subjected to bending and axial force is presented. Timber is conservatively modelled as an elasto-brittle material, whereas metal connections (hold-downs and angle brackets) are modelled with an elasto-plastic behavior. The reaction force in each connection is iteratively calculated by varying the position of the neutral axis at the base of the wall using a simple algorithm that was implemented first in a purposely developed spreadsheet, and then into a purposely developed software. This method is based on the evaluation of five different failure mechanisms at ultimate limit state, starting from the fully tensioned wall to the fully compressed one, similarly to reinforced concrete (RC) section design. By setting the mechanical properties of timber and metal connections and the geometry of the CLT panel, the algorithm calculates, for every axial load value, the ultimate resisting moment of the entire wall and the position of the neutral axis. The procedure mainly applies to platform-type structures with holddowns and angle brackets connections at the base of the wall and rocking mechanism as the prevalent way of dissipation. This method allows the designer to have information on the rocking capacity of the system and on the failure mechanism for a given distribution of external loads. The proposed method was validated on the results of FE analyses using SAP2000 and ABAQUS showing acceptable accuracy.
In this paper, a simplified non-linear procedure for seismic design of CLT (cross-laminated timber) wall systems is presented. The proposed method considers both axial force and bending moment applied on the wall systems as result of applied loads. Timber is modelled as an elastic-brittle material, whereas metal connections (hold-downs and angle brackets) are modelled with an elastic-plastic behaviour. The reaction force in each connection is iteratively calculated by varying the position of the neutral axis at the base of the wall using a simple algorithm that has been implemented in a purposely-developed software. This method is based on the evaluation of five different failure mechanisms at ultimate limit state similarly to reinforced concrete (RC) rectangular section design. By setting the mechanical properties of timber and metal connections, and the geometry of the CLT panel, the algorithm calculates, for every possible axial load, the position of the neutral axis and the ultimate resisting moment of the system. Furthermore, this method also allows the designer to have an indication on the failure mechanism of the wall.