This study develops an end-to-end workflow, from laboratory measurements to Eulerian–Eulerian two-phase simulations with SedFoam, to investigate bed erosion in free-surface open-channel flow over a deformable granular bed. Experiments were conducted with a calibrated non-cohesive deposit of epoxy-coated spherical beads under steady, fully turbulent, subcritical conditions. Particle Image Velocimetry provided mean-flow and turbulence data, while a 3D camera workflow supplied bed-elevation fields and time-resolved maps of sediment rearrangement. These datasets were used to constrain a staged numerical strategy in which single-phase hydrodynamics were first reproduced and then extended to live-bed morphodynamics. Validation over a rigid bed showed that the 2006 k–ω closure, combined with a rough-wall treatment, reproduced the measured mean-velocity profiles and provided acceptable turbulent kinetic energy levels, yielding dynamically consistent near-bed shear conditions. In live-bed conditions, the simulations reproduced the streamwise organization of scour and deposition, predicted cumulative erosion rates of the correct order of magnitude, and captured bedform migration consistent with time-resolved bed reconstructions. The numerical results were compared with repeated experiments while accounting for run-to-run variability and the metrological limits of the 3D camera. This work proposes a transferable experimental–numerical methodology for assessing the predictive capability of live-bed morphodynamic simulations, in which hydraulic characterization, three-dimensional bed monitoring, erosion/deposition metrics, and repeated experiments are combined within a common comparison procedure.
Ennazii et al. (Mon,) studied this question.