This study investigated the vibration serviceability of timber structures with dowel-type connections. It addressed the use of such connections in cutting-edge timber structures such as multi-storey buildings and long-span bridges, in which the light weight and flexibility of the structure make it possible that vibration induced by dynamic forces such as wind or footfall may cause discomfort to occupants or users of the structure, or otherwise impair its intended use. The nature of the oscillating force imposed on connections by this form of vibration was defined based on literature review and the use of established mathematical models. This allowed the appropriate cyclic load to be applied in experimental work on the most basic component of a dowel-type connection: a steel dowel embedding into a block of timber. A model for the stiffness of the timber in embedment under this cyclic load was developed based on an elastic stress function, which could then be used as the basis of a model for a complete connector. Nonlinear and time-dependent behaviour was also observed in embedment, and a simple rheological model incorporating elastic, viscoelastic and plastic elements was fitted to the measured response to cyclic load. Observations of the embedment response of the timber were then used to explain features of the behaviour of complete single- and multiple-dowel connections under cyclic load representative of in-service vibration. Complete portal frames and cantilever beams were tested under cyclic load, and a design method was derived for predicting the stiffness of such structures, using analytical equations based on the model for embedment behaviour. In each cyclic load test the energy dissipation in the specimen, which contributes to the damping in a complete structure, was measured. The analytical model was used to predict frictional energy dissipation in embedment, which was shown to make a significant contribution to damping in single-dowel connections. Based on the experimental results and analysis, several defining aspects of the dynamic response of the complete structures, such as a reduction of natural frequency with increased amplitude of applied load, were related to the observed and modelled embedment behaviour of the connections.
Cross laminated timber (CLT) panels were manufactured and tested to assess their time dependent behaviour. This study is intended to help guide the development of an appropriate test method and acceptance criteria to account for duration of load and creep effects in the design of structures using CLT.
Nine CLT panels of different qualities and using different wood species combinations were manufactured at a pre-commercial pilot plant out of local wood species. The CLT panels manufactured in this study were pressed at about 54% lower pressure than the minimum vertical pressure specified by the adhesive manufacturer due to a limitation of the press, so the CLT panels are viewed as a simulated defective sample, which may occur in a production environment due to material- or process-related issues.
ASTM D6815-09 provides specifications for evaluation of duration of load and creep effects of wood and wood-based products. The standard was designed to accommodate wood products that can be easily sampled, handled, and tested under load for minimum 90 days and up to 120 days. The standard requires a minimum sample size of 28 specimens. Because of its large dimensions, CLT products are not feasible for experiments requiring such large sample sizes. However, the findings of this study revealed potential for some of the acceptance criteria in ASTM D6815-09 to be applied to CLT products. The CLT billets in this study were assessed in accordance to the creep rate, fractional deflection, and creep recovery criteria in ASTM D6815-09 standard. All CLT billets tested in this study showed (1) decreasing creep rates after 90/120 days of loading, (2) fractional deflections less than 2.0 after 90-day loading, and (3) higher creep recovery than 20% after 30 days of unloading, as required by ASTM D6815-09. A single replicate billet was used per CLT configuration instead of the minimum sample size required by the standard which may have an effect on the findings.