Stochastic simulations of headed-stud shear connections in profiled composite floors

In this study, recycled aggregate concrete (RAC) is applied on the top of profiled
steel decking to form composite slabs, working together with steel beams via
commonly used headed-stud shear connections. The authors used advanced finite
element (FE) simulations to reveal the load-bearing behaviour (i.e., load capacity,
slip modulus, deformation capacity, and failure modes) of headed-stud RAC
connections. Prior to simulations, the FE model is verified by available shear tests
on headed-stud connections using natural aggregate concrete (NAC).
Deterministically, the headed-stud RAC and NAC connections have similar loadbearing
behaviour, providing that there is negligible difference in the strength of RAC
and NAC. However, the high variability of mechanical properties of RAC is an
essential safety concern for structural applications. To address the influence of
uncertainty (high variability) of RAC mechanical properties on the resistance of
headed-stud shear connections, stochastic functions are established with polynomial
chaos expansion technique based on results of advanced FE simulations. The
stochastic functions are used to generate a database of resistance of a pool of
headed-stud connection models containing random properties of RAC. With the
database of headed-stud RAC connections, one can perform reliability analyses for
calibrating partial safety factors in the future design model.