On modeling subgrid-scale macro-structures in narrow twisted channels. (January 2020)

• Record Type: Journal Article
• Authors: Li, Zhi; Hodges, Ben R.
• Is Part Of: Advances in water resources. Volume 135(2020)
• Publication Date: 2020-01
• Subjects: Subgrid topography model -- Grid alignment -- Narrow twisted channel -- Subgrid macro-structures -- Topographic dissipation; Hydrology -- Periodicals; Hydrodynamics -- Periodicals; Hydraulic engineering -- Periodicals; 551.48
• DOI: 10.1016/j.advwatres.2019.103465
• Languages: English
• ISSNs: 0309-1708

Abstract

Highlights: Subgrid-scale macro-structures enhance topographic dispersion. Reducing cell volume and face areas can model effects of macro-structures. Alignment (or not) between grid lines and channel direction affects model results. Quantification of macro-structure geometry is a key task for future study. Abstract: Porosity-based subgrid topography models often fail to capture the effects of subgrid-scale topographic features in the interior of coarse grid cells. Existing approaches that modify bottom roughness or a drag coefficient are inadequate for macro-structures (large emergent or submerged obstacles) in subgrid-scale narrow twisted channels. Such structures partially block the cross-sectional area and provide enhanced topographic dissipation – effects that are not well represented by a drag coefficient that scales on a coarse-grid cell-averaged velocity and the cell volume. The relative alignment between mesh and flow further complicates this problem as it makes the subgrid model sensitive to mesh design. In the present study, three new approaches for simulating subgrid-scale macro-structures in narrow channels are proposed. The interior partial-blocking effect of structures is modeled as reduction of grid face-area. The sheltering of flow volumes around obstacles, which leads to topographic dissipation, is modeled by reducing the cell volume in the momentum equation (only). A mesh-shift procedure is designed to optimize mesh alignment for identifiable subgrid features. Combining the three subgrid methods improves the approximation of surface elevation and in-channel flow rate with a coarse-grid model. Tests are conducted for channelized flow using both synthetic domains and real marsh topography. The new methods reduce the overall mesh dependency of the subgrid model and provides stronger physical connection between effects of macro-structures and their geometry at coarse grid scales.