The aim of this study was to investigate the load-bearing capacity and stiffness of mechanical and adhesively bonded connections in timber structures with gusset plates made of birch plywood, as a preliminary study for potential future use in long-span timber structures which could reduce the environmental impact of the building sector by substituting steel with wood products in moment-resisting connections. Birch plywood was chosen for this study due to its high bending strength in comparison to plywood made of softwoods and its availability on the market.

The specimens were made up of two GL28cs glulam beams jointed at mid-span by two birch plywood plates on each side. For the adhesively bonded connections, two different bonded areas were tested to investigate a possible influence of a size effect. Furthermore, two different plywood face-grain angles of 0 ° and 22.5 ° were tested for each test configuration, to investigate the effect that the face-grain angle of the plywood has on the connections. Both types of connections were subjected to a four-point flexural test with a universal testing machine, to create a pure bending moment at mid-span of the connections. During the tests, vertical displacement and rotation at mid-span was measured in order to calculate the stiffness of the joints. The experimental results were thereafter compared to predicted failure loads obtained from analytical models and the mechanical joint’s failure mode simulated by a numerical model in Abaqus.

The experimental results showed that the stiffness of the specimens with a plywood face-grain angle of 0 ° were significantly greater than the specimens with 22.5 ° face-grain angle. Adhesively bonded joints also showed high potential for use in moment-resisting connections, being approximately twice as stiff as the mechanical joints. Additionally, allowing the 2-component polyurethane (2CPUR) adhesive to rest for 10 minutes after mixing while exposed to air, for the purpose of increasing the adhesive’s viscosity, decreased adhesive failure in the joints and resulted in stronger adhesion and failure modes of almost 100 % wood failure in all specimens. Furthermore, there is a size effect regarding total load capacity when comparing the adhesively bonded areas, as the total load capacity increased more than the bonded area. However, the same could not be said about shear stress capacity where the results were inconclusive.

The analytical results showed that the predicted failure loads were very conservative compared to the actual performance of the joints. The mechanical joints had a capacity that was much greater than predicted, which may be due to an unaccounted strength contribution through the "rope effect" when the screws were in the plastic state. The analytical models for the prediction of load-bearing capacity of adhesively bonded joints also need further development.

Future work on this topic should include investigating the influence of plywood face-grain angles on rotational stiffness and load-bearing capacity in non-parallel connections, as well as more research on the size effect in adhesively bonded joints and the development of analytical models for these types of connections. Large scale tests of portal frame corners with gusset plates made of birch plywood would also be beneficial to research