"Composite wood-concrete beams using utility poles: time-dependent behavior"

This paper describes the behavior of wood-concrete composite beams in which the wood layer is composed of recycled utility poles. Two full-scale laboratory specimens of 7,500-mm (295-in.) span were constructed, with the concrete slab cast on the timber poles left unshored. The connection was obtained by cutting six notches in the timber poles. A threaded steel dowel is placed in each notch extending through both layers with a nut on the bottom of the beam that can be tightened once the concrete has cured. Since the material in each layer exhibits a time-dependent response, it is necessary to examine the effects of sustained loading. After 28 days of initial curing of the concrete layer during which the midspan deflection was monitored, both beams were subjected to the service load to estimate the composite efficiency. The beams were very stiff and values of 96% and 98% were obtained, respectively, demonstrating the high efficiency of the notched connection system used. A comparison was made with results obtained in the literature for ramp load tests of two similar specimens, which had exhibited composite efficiencies of about 94% and 96%, respectively. A sustained load of approximately 11% of the estimated static ultimate load capacity was then applied to one of the specimens, which was monitored over time. The applied load resulted in a midspan moment of 24 kN•m (212 kip•in) while the calculated ultimate capacity corresponded to a midspan moment of 226.5 kN•m (2,005 kip•in). The final deflection was found to increase to 44.4 mm (1.75 in.) after 256 days, which was more than twice the initial elastic deflection owing to the application of the dead and live loads. The deflection attributable to the dead load was 19.1 mm (0.75 in.), and the deflection attributable to the live load was 2.54 mm (0.10 in.). An existing one-dimensional (1D) finite-element program was utilized to predict the time-dependent deflection at the end of a 50-year service life. The comparison with experimental results showed good approximation, particularly for the midspan deflection. The software predicted a 54-mm (2.14-in.) midspan deflection at the end of a 50-year service life, corresponding to 1/139 of the span. This was sufficiently high to suggest that either larger cross-section and/or precambering of the wood layer could be needed when the serviceability limit state of maximum deflection in the long-term is a required performance criterion.