Title
3-D dam break flow simulations in simplified and complex domains
Date Issued
01 March 2020
Access level
open access
Resource Type
journal article
Author(s)
Constantinescu G.
The University of Iowa
Publisher(s)
Elsevier Ltd
Abstract
This paper presents a 3-D, non-hydrostatic, Reynolds-Averaged Navier-Stokes (RANS) model using the volume of fluid (VOF) approach to simulate dam-break flows. Good agreement is observed between the 3-D model predictions and results of dam-break experiments performed in the laboratory. The 3-D model is then applied to predict flood-wave propagation induced by the sudden failure of two flood-protection dams in Iowa, USA. Results are also compared with predictions of 2-D, hydrostatic, depth-averaged models. The 2-D model simulations using the precalibrated values of the Manning's coefficients underpredict the speed of propagation of the flood wave and the area inundated by the flood compared to the 3-D model predictions. A methodology is presented to recalibrate the 2-D model which improves the agreement with the 3-D model predictions. Simulation results also show that strong 3-D effects are generated in regions of strong curvature of the river channel, near sudden constrictions and obstacles, and during the times the mean flow direction varies significantly over the flow depth. Such 3-D effects cannot be captured by the 2-D model even after recalibration, pointing toward the need to use 3-D models for detailed flood mapping.
Volume
137
Language
English
OCDE Knowledge area
Ingeniería ambiental y geológica
Geotecnia
Subjects
Scopus EID
2-s2.0-85078153674
Source
Advances in Water Resources
ISSN of the container
03091708
Sponsor(s)
The authors would like to acknowledge the Iowa Flood Center and Prof. W. Krajewski for supporting this study and providing guidance on many aspects of the reported research. The authors would also like to acknowledge high-performance computing support (project DD-2015-GEO117) from the Oak Ridge Leadership Computing Facility (OLCF) and, in particular, Dr. J. Wells from the National Center for Computational Science at the Oak Ridge National Laboratory as well as from the High-Performance Computing Center at the University of Iowa.
Sources of information:
Directorio de Producción Científica
Scopus