Seismic performance of dowel-type connections in tall timber buildings.

Abstract

The PhD research reported in this thesis explores the seismic performance of dowel-type connections in (tall) timber buildings. The research was in part motivated by the fact that connection design in design codes as well as ductility and overstrength definitions are based on small and medium-scale connection tests, and therefore these calculation rules needed to be verified for large-scale hold-downs used in tall timber buildings. The research consisted of an extensive experimental program and the development of analytical models. The experiments comprised of embedment tests, monotonic and cyclic small- and large-scale tests of dowelled connections in Cross Laminated Timber (CLT) and Laminated Veneer Lumber (LVL), as well as material tests of small clear sawn timber and LVL specimens. It was found that the embedment strength formula given in the CLT Handbook, fh,k = 0.031(1 − 0.015d)ρ1.16, is generally applicable for New Zealand CLT made out of radiata pine within the bounds of 0.95 ≤ tI/ t⊥ ≤ 2.75, were t is the layer thickness parallel and perpendicular to the outer layer. The small-scale and large-scale connection behavior was evaluated in terms of strength, stiffness, ductility, energy dissipation, and overstrength. It was found that medium to high ductility can be achieved provided that brittle failure is prevented by ensuring the governing brittle mode is protected from the overstrength of the ductile capacity. Furthermore, ductility can be improved by increasing the dowel row spacing and dowel end distance, and by substituting the outer CLT layers with LVL. Finally, the calculation rules provided in Eurocode 5 were applicable to large-scale connection design.

Particle Tracking Velocimetry (PTV), a quantitative field measuring technique originally designed to track individual particles in fluid flows, was applied for the first time in the context of structural timber engineering. Paint particles were applied to the timber surface and their movements were tracked in order to derive strain and displacement fields. It was possible to capture dowel displacements, dowel embedment, specimen displacements and rotations, as well as crack growth and brittle failure.

In terms of theoretical contribution, models for brittle failure of dowel-type connections and mode cross-over from a ductile to brittle response were extended and refined. Furthermore, a component based analytical method to derive overstrength of hold-down-type steel-to-timber connections based on material properties was developed. Design values for the overstrength sub-components were established by means of parameter studies. Lastly, generally applicable overstrength factors for dowelled connections with a slotted-in steel plate were derived for New Zealand sawn timber / CLT and New Zealand LVL as γRd,th,sawn = 1.7, and γRd,th,LVL = 1.5, respectively.