Grid Independence Studies of Eddy-Resolving Models at the Scale of a River Reach Along a Transect in Marble Canyon of the Colorado River, Arizona
Abstract
This study presented a novel approach for evaluating the impact of changes in spatial resolution on field-scale models to identify turbulent flow patterns in hydraulic features in canyon-bound rivers. The methodological framework incorporated seven different computational domain resolutions to analyze the sensitivity of field-scale models to spatial resolution changes by employing techniques such as turbulent kinetic energy spectrum, spatiotemporal analysis of eddy structures, divergent Kullback-Leibler (KL) divergence, Nash-Sutcliffe (NS) model efficiency coefficient, wavelet power spectrum, and GCI calculation. Statistical and physical-based skill metrics are employed to quantifying information loss to evaluate a set of domain resolutions. Grid independence studies have become relevant tools for understanding the effect of domain resolution in the simulation of anisotropic turbulence at the river-reach scale while using LES or DES models. The primary goal of a grid independence study is to estimate an optimal grid by evaluating the flow quantities in different domain resolution conditions without generating a difference in the numerical results. In this research, a grid independence study has been conducted along a transect of the Colorado River in Marble Canyon, Arizona, to improve the performance of parallelized physics-based models, minimize computational expenses, and maximize model accuracy. The model employs the Detached Eddy Simulation (DES) technique, which resolves turbulence structures larger than the grid spacing in the interior of the flow. The Boundary Condition (BC) in the bed is non-slip and integrates the rough wall extension of the Spallart-Allmaras model in the cells near the bed. The rough wall function is built into the OpenFOAM environment and uses the velocity in the first grid cell near the bed to calculate the turbulent viscosity along the wall. The medium resolution was the minimum grid size required to obtain accurate simulation results for capturing flow behavior and hydraulic characteristics, including secondary flows, return currents, shear layers, and primary and secondary eddies. The study also showed that medium resolutions captured smaller scales of turbulence and provided a more accurate data representation of the Turbulent Kinetic Energy (TKE) spectrum than the benchmark case scenario. The RMSE surface analysis identified areas of the flow field that were particularly sensitive to grid sizes.
Subject Area
Engineering|Hydrologic sciences|Geomorphology|Environmental science
Recommended Citation
España Urresty, Rosa Elena, "Grid Independence Studies of Eddy-Resolving Models at the Scale of a River Reach Along a Transect in Marble Canyon of the Colorado River, Arizona" (2023). ETD Collection for University of Texas, El Paso. AAI30524963.
https://scholarworks.utep.edu/dissertations/AAI30524963