Dam-Break Modeling: LBM as the Way towards Fully 3D, Large-Scale Applications

Abstract

Free-surface flows play an important role in many environmental hydraulic problems. Owing to their characteristic large spatial and temporal scales, reduced complexity models, such as depth-integrated shallow water (DI-SW) models, have so far been largely preferred to three-dimensional (3D) approaches. Even if they can reproduce water levels satisfactorily, DI-SW models do not natively provide flow velocity at the bed, which is pivotal for estimating scour effects induced by impulsive flows, such as those resulting from dam breaks. However, resort to fully 3D modeling is often impractical due to the high computational demand required; nonetheless, recent works have shown impressive performance of the lattice Boltzmann method (LBM) compared to classical numerical methods while proving to be as accurate in many fields, suggesting LBM makes such computations affordable on desktop computers. So far, a comprehensive assessment of a LBM-based, fully 3D, free-surface model is still lacking. In this work, a LBM model is compared to data from a large series of dam-break-flow experiments, in terms of water levels and, where available, flow velocities. Comparisons with available DI-SW models are also carried out. Results show that the LBM model can correctly simulate both water levels and velocity distributions, the latter often being inaccurately reproduced by DI-SW models. To provide guidelines for the correct modeling of such flows, considerations are also given to opportunities to use either no-slip or free-slip boundary conditions, the latter being preferable if underresolved boundary layers are present.