This paper describes the applicability of a multi-level grid type turbulent shallow flow model (KMR: Kinematic Mesh Reconstruction) to an unsteady open channel flow around a bridge pier. The KMR model employed a two-dimensional depth-averaged URANS (Unsteady Reynolds Averaged Navier Stokes) type shallow flow equations. As a turbulence model, a second-order non-linear 0-equation model proposed by Kimura et al (2009)1) is adopted. The turbulence structure in a shallow open channel flow is characterized by vortex merging and up-cascading. To predict efficiently a large spatial difference of the turbulence scale, computational grid is divided and merged automatically and dynamically at each time step. The model is applied to a shallow flow around a square bridge pier model and the accuracy of computational results on vortices behavior, velocity profile and water surface elevation is examined through the comparison with a previous experimental result. Numerical results showed that the present model could capture excellently the fundamental aspects of the flow around a bridge pier with less CPU time than the fixed grid computation.
