Stress-laminated-timber bridge decks subjected to ultimate loads

Licentiate thesis, 2011

Stress-laminated-timber (SLT) bridge decks are a satisfactory alternative to conventional short-span bridges in terms of cost and performance. SLT decks are made from several deck laminations positioned side by side. The laminations in the deck are normally timber or glulam beams that are stressed together using high-strength steel rods positioned in pre-drilled holes perpendicular to the length of the laminations. A concentrated load is distributed from the loaded lamination onto the adjacent laminations due to the resisting friction between the stressed laminations. The pre-stress between the laminations should be of such a magnitude that movements between the laminations are prevented. The general assumption regarding the design of stress-laminated-timber (SLT) bridge decks is that the structural behaviour is linear in both the serviceability-limit state (SLS) and the ultimate-limit state (ULS). In this thesis, the validity of this assumption is investigated. Several full-scale tests have been conducted in order to determine the structural response and performance of SLT decks subjected to high concentrated loads causing the failure of the decks. Several complementary tests have also been conducted in order to investigate the behaviour of specific details in SLT decks. Several approaches to the design of SLT decks are currently in use. Both “hand calculation” methods and numerical finite element methods are commonly used. The hand calculation methods are generally based on the “equivalent beam theory”. In this method, the three-dimensional deck is replaced by a two-dimensional beam which has a width corresponding to the load distribution in the transverse direction. Numerical FE analyses are normally conducted using three-dimensional shells. The shell is assigned orthotropic mechanical material parameters in order to include the load distribution in the transverse direction. Tests have shown that the load-deflection relationship of the decks was non-linear when subjected to eccentric loads. Interlaminar slip occurred between the laminations in the deck. These small movements caused significant stress redistribution in the deck. The interlaminar slip was dependent on the pre-stress level. The horizontal interlaminar slip occurred at low loads of approximately 10 20% of the ultimate load capacity. None of the linear design models proved sufficient to include these non-linear effects.