CFD modeling of fish passage in large culvert and assistance for culvert design with fish passage

Culverts have been traditionally designed and constructed to be hydraulically efficient for conveying flood flows. The U.S. Federal Highway Administration is currently focused on modifying traditional design methods to achieve both the hydraulic and biological efficiencies. Current information and culvert manufacturers typically addresses flow magnitudes in the order of the bankfull stage or greater. It is time to get more information about the hydraulics of low flow in large culverts.;This dissertation is specifically focused on the Computational Fluid Dynamics (CFD) study for circular culvert flow. The simulations were performed by solving the three-dimensional Reynolds averaged Navier-Stokes equations with the k -- epsilon turbulence model for steady-state flow. The computational efficiency was improved by the development from the multi-phase model to the truncated single phase model with cyclic boundary. The supra-ball model performed well in representing the roughness from gravel bed.;The CFD models were verified by comparing with data (both PIV and ADV) from experiments conducted at the TFHRC. A statistical analysis of the error of the CFD modeling data fit to the experiments was performed. Scale validation of the truncated single phase CFD culvert model was performed and the CFD culvert model was applied into prototype. In order to evaluate the ability of fish to traverse corrugated metal culverts, the local average velocity was determined in vertical strips from the wall to the centerline of the culvert.;After successful simulation of culvert flows, the method was further extended for mathematical equations to predict the flow distributions. Chiu's equations were used to predict 2-D velocity distributions over the cross sections. Velocity distributions were obtained using the non-linear least-squares regression method. Overall, the semi-empirical equations did well in predicting 2-D velocity distribution on the culvert cross section with the exception of under-predict value of averaged velocity. However, the depth-averaged velocity curve is acceptable for engineering applications.