Long-term behavior of continuous composite slabs made with 100% fine and coarse recycled aggregate

•Long-term behavior of continuous composite slabs with recycled aggregate concrete.•Influence of non-uniform shrinkage on crack width of continuous composite slabs.•Influence of fine and coarse recycled aggregate on continuous composite slabs.•Modeling the long-term behaviors of composite slabs. A composite slab is an efficient modern structural member. Recycled fine and coarse aggregates are alternative green materials for the depleting natural sand and natural gravels in the composite slabs. In recent years, the nonuniform shrinkage occurring in single-side-exposed concrete has been discovered to significantly affect the long-term behaviors of composite slabs, which have been included in newly published codes. Applying recycled aggregate concrete (RAC) would further increase the effect of nonuniform shrinkage on the long-term behaviors of composite slabs. However, most studies, with RAC or not, have only focused on simply- supported composite slabs. For continuous composite slabs (CCSs), in addition to increasing deflections, a nonuniform shrinkage would induce a larger negative bending moment, which would further increase the crack width at the interior support region. Hitherto, no experiment has been conducted regarding the time-dependent crack width of CCSs and CCSs with RAC. Therefore, the aim of this study is to investigate the effects of nonuniform shrinkage and RAC on the deflections, strains, and cracks of CCSs. Hence, a 500-d experiment was conducted on three full-scale specimens. Finally, a finite element (FE) model was established and benchmarked against the test results of this study and those reported in the literature. The results indicated that (1) nonuniform shrinkage significantly increased the crack width at the negative bending moment region; (2) the use of RAC resulted in a more significant effect of the nonuniform, e.g. mid-span deflections, soffit strains, and crack widths increased by 40%, 64%, and 38%, respectively; and (3) the FE model can predict the long-term behaviors of CCSs well.