Best Practices in Developing and Reviewing CFD Models for Spillways

ABSTRACT ONLY - Some dams have unique and complex spillway geometries that make the hydraulic design and analysis of those structures complex. For such dams, Computational Fluid Dynamics (CFD) modeling, a sophisticated approach to numerically solve hydraulics problems, is emerging as a great tool. As technology has been advancing and high performing computers are becoming more and more accessible, CFD modeling has become a powerful, economical, and therefore, popular tool in the hands of modelers and designers. While CFD usage is becoming more prevalent in the dam safety industry, many professionals without much prior experience with this tool may be tasked with developing and/or reviewing CFD models. This presentation aims to provide guidance on best practices for both parties; those developing and those reviewing CFD models. For modelers, it is important to know the significance of key model input parameters that impact the results and when sensitivity analyses to these parameters may be necessary. Such parameters, whose appropriateness is important to achieve reliable modeling, include three-dimensional model geometry, simulation duration, model boundary location, turbulence approximation, mesh cell size, and surface roughness. Modelers must also understand how to extract and report the output data necessary to verify and gain confidence in the model results. Reviewers of CFD models should have a fundamental knowledge of modeling objectives, fluid dynamics equations, turbulence models, numerical methods, mesh quality, model validation, and model parameters. To provide a thorough and informed review, it is crucial that modeling approaches, major assumptions, model set up inputs, sources of all boundary conditions data, mesh details, key geometry dimensions, design drawings, and model results are adequately documented. The authors will define and discuss each of the aforementioned aspects of CFD modeling using recent project examples from the perspective of both the modeler and the reviewer.