A new laboratory benchmark for validating non-hydrostatic depth-averaged models in rapidly varied free-surface flows

This work presents a new laboratory benchmark for the validation of non-hydrostatic free-surface-flow models under rapidly varied open-channel conditions. The test case consists of a steady free-surface flow through a non-uniform channel geometry featuring a marked contraction and subsequent expansion in plan, combined with abrupt variations in bed elevation. This configuration was specifically designed to induce rapid accelerations and decelerations of the flow in both the longitudinal and vertical directions, transitions between subcritical and supercritical regimes, and significant deviations from the hydrostatic pressure distribution. The benchmark is documented for two discharges, Q = 12 l/s and Q = 15 l/s, through a coordinated set of measurements, including free-surface elevations, bed and sidewall pressure distributions from 42 piezometric taps, and Particle Image Velocimetry velocity fields in a vertical plane within the contracted reach. The experimental dataset is then used to assess a discontinuous Galerkin finite-element model solving the two-dimensional Vertically Averaged and Moment (VAM) equations. The comparison shows that the VAM approach reproduces the measured free-surface profiles, bed and sidewall pressure distributions, and streamwise and vertical velocity profiles more accurately than the Shallow Water equations model for both discharges. The proposed dataset, therefore, provides a challenging, well-documented experimental benchmark for the validation of non-hydrostatic depth-averaged models and fully three-dimensional free-surface-flow solvers.