A New Approach to Streambed Modeling and Simulation Using CFD

Realistic three-dimensional computer models of natural streams and rivers have become a priority for many civil works researchers, particularly in light of the exorbitant costs associated with typical stream restoration. Fortunately, the rapid advances and availability of high performance computing resources along with the increased sophistication of both in-house and commercial software have made the creation of such models significantly more tenable. In light of these needs, the objectives for this research include: (1) the creation of realistic representations of naturally occurring streambeds from potentially coarse sets of field measurements; (2) the demonstrated capability to freely deform the streambed surface as well as embed complex objects within the surface (rocks, fallen trees, etc.); (3) the ability to successfully mesh the surface and its surrounding volume in accordance with established mesh quality criteria; and (4) the ability to obtain sufficiently resolved flow field solutions utilizing HPC resources. Beginning with a coarse set of field data measurements taken from one of four study sites along a 1.5-mile stretch along the Robinson restoration project of the Merced River, California, we will demonstrate, for each objective stated, the respective challenges, solution strategies, and resulting outcomes. Flow field solutions are conducted using parallelized finite element/volume solvers.