Axially loaded screw joints in wooden structures exposed to fire : Fire tests in furnace and numerical calculations in TASEF of different screw mounting variations

Abstract: To achieve sufficient stability of a structure in a fire scenario, a thorough design of joints in wooden structures are as important as the design of the construction itself. As an earlier study points out, there are no existing design method available to determine the load capacity of an axially loaded screw joint in wooden structures, which complicate the process of a thorough design of the joint in a fire scenario.This research has been made with the aim to establish a sufficient design approach in terms of the most effective mounting of the screw inside a wooden joint with respect to load capacity in a fire scenario. A numerical model has been created in TASEF to evaluate if this type of model can predict the temperatures inside a joint, and further be applicable in the design phase of the joints regarding to the load capacity´s temperature dependency.The objective of this research has been the study of three different types of screw joints exposed to axial load, with two different screw lengths. Laboratory work has been made, where four reference tests were performed to gain reference load capacities to later be compared with the load capacities of twelve fire tests. The fire tests were conducted with fire exposure of 30 and 60 min, where the load capacity and the temperatures inside the joint where measured. Temperature models were created in TASEF to represent each fire test, where the temperatures inside the joints where calculated and evaluated.The results showed that the most effective screw mounting with respect to remaining load capacity after fire exposure was case 3, where a protective dowel was added to the countersunk screw. This type of mounting had the greatest remaining fastening length because the charring did not affect this length in the same extent as for the other two cases. The longer screw had a percentual higher remaining load capacity than the shorter screw. This can also be described by the relationship between the fastening length and the load capacity.In specific points, the numerical model did not correspond to the measured temperatures. One explanation for this scenario can be the fact that the material properties of the screw used in the fire tests were not equal to the properties used in TASEF. This may have affected the conductive heat flux through the joint, leaving different temperatures in the two measuring points. However, a good correlation between the charring of the test specimens and the temperature distribution achieved from the numerical model in TASEF exists and this implies that this model can be used for predicting the general temperature distribution inside a fire exposed joint.