The restoration and preservation of historic timber structures depends on the integrity of the structural members. Structural properties need to be assessed for the individual members and for the connections between those members. An array of methods has been developed to non-destructively determine mechanical properties of wood in-situ. However there is no widely accepted method for non- destructively determining the structural properties of connections in timber structures.
This research work field tested a rapid method that had been developed and validated in laboratory work. The work is based on the theory that the natural frequency of vibration of the beam depends on the stiffness of the beam and the stiffness of the connections at the end of the beam. Just as changing the tension in a guitar string changes its frequency, so does change the degree of rotational restraint at the end of the beam. The frequency of vibration of the beam can determine how rigid (stiff) or pinned (loose) the joints in a building are, indicating connection differences between repetitive members and providing the essential steps in performing a structural analysis of the building.
Six buildings of varying age, size, archetype, and connection type were studied. In each of these buildings, critical members were identified that had a clear span that allowed them to vibrate freely. For each of these members, the basic geometric properties were measured, and the Modulus of Elasticity was estimated based on species and grade identification. The member was then instrumented with an accelerometer and allowed to vibrate freely. The time history of the vibration was recorded using a laptop computer. The corresponding frequency spectrum was analyzed, and the fundamental frequency identified. Based on a closed form solution, the frequency and other measured properties could be used to determine the rotational stiffness of the supports.
The results from field testing provided a number of significant findings. First, the wide-spread assumption that timber joints have insignificant rotational stiffness (α < 1)is not valid. 70% of the 20 joints tested had stiffness values that classified them as semi-rigid (1 < α < 100). Second, the finding
from lab work was confirmed: the ability to accurately predict the stiffness improves as the joint stiffness increases from pinned to semi-rigid (Percent error of 84% for α<1, and 41% for 1< α <100). Finally, the method showed anecdotal evidence that joints with lower measured stiffness had visual defects evident upon inspection.