Development and testing of a numerical code for treatment of complex river channel topography in three-dimensional CFD models with structured grids

The potential for using high-resolution numerical modelling to understand flow over complex rough heterogeneous surfaces has yet to be fully realized, largely due to problems of designing numerically stable meshes for use with complex bed geometries. Representation of such geometries has tended to rely upon sub-grid scale treatments involving roughness parameterization within numerical schemes. This paper develops, verifies and validates a technique for dealing with the representation of complex bed geometries within a basic numerical scheme. The method is based upon a numerical porosity treatment, which uses a structured grid but specifies cell porosities to block out bottom topography (P = 1 for cells that are all water, P = 0 for cells that are all bed and 0 < P < 1 for partly blocked cells) with appropriate drag terms introduced into the momentum equations. This method has the distinct advantages of minimizing grid distortion and increasing the ease of numerical solution. A validation exercise is presented that is based upon two separate flume experiments using digital particle imaging velocimetry to collect velocity measurements of (i) flow around a cube; and (ii) a simulation of a simplified representation of bed roughness. These results are considered against simulations based upon previous treatments of complex topography, notably in relation to roughness parameterization.