Reconsidering Composite Action on Strength of Wood Roof Systems

Structural aspects of the installation of photovoltaic (PV) modules on existing wood roof framing systems are considered. Often, existing roofs are inadequate to resist even current minimum design loads, much less the extra weight of PV modules. Rather than requiring expensive retrofits, it is desired to bring these under-strength roofs into compliance with current codes by reconsidering composite action, produced by the joist-sheathing interaction of the assembly, on the bending strength of joists. The effect of composite action can be quantified by considering a composite T-beam, with the wood joist being the web and the structural sheathing being the flange. Although this effect has already been considered in the past (under different assumptions and different models), it is hypothesized that an increase in strength can still be achieved even when using a discontinuous flange (due to gaps between sheathing panels), and non-rigid connections like a nailed joint. Laboratory experiments using several 2”x 4” bare wood joists and 2”x4” joists along with 7/16”-thick OSB panels attached to their tops (using nailed and glued-and-nailed connections) were performed. The experiments showed bending strength increases between 18 to 60% in T-beams, when compared to the bending strength of bare joists. Tests also showed that the strength of the T-beam varies as a function of the random location, with respect to the location of a gap, of a weak knot with lower strength that initiates failure in the joist. The behavior and mechanical properties of the connection are also studied and compared with existing literature. Analytical and finite element analyses of the composite T-beam subjected to constant moment, show nonlinear behavior of the nailed connection when the failure moment is applied to the composite T-beam, hence, developing Nonlinear Partial Composite Action (NPCA). A small decrease in the maximum tensile stress at the bottom of the joist is obtained at sections away from the location of the gap due to NPCA. Also, finite element analysis shows that even at the location of the gap, the tensile stress is slightly reduced due to local effects in the joist produced by the local nail forces. A Monte Carlo simulation is utilized to randomly vary the location of knots along the span of the T-beam and conservatively evaluate the effect that NPCA has in the tensile stress at each knot. Slight increases in the mean bending strength of the joist are predicted, depending upon the 5% exclusion limit bending strength value for clear wood and the stress grade of the species. However, these predictions of strength increase are small and do not match the 18 to 60% observed increases in the laboratory experiments. Perhaps the number of laboratory tests was insufficient to produce a statistically significant data set. Finally, generic, easy to install and effective procedures for retrofitting roof joists to enhance the beneficial effects of NPCA are proposed to justify increased roof strength to allow for the loads from the PV modules.Sandia National LaboratoriesCivil EngineeringMastersUniversity of New Mexico. Dept. of Civil EngineeringGerstle, WalterMaji, ArupRoss, TimothyDwyer, Stephe