3.6.32D/1D Models
Prior to Build 2005-05-AN, 2D/1D models use the same timestep in both 1D and 2D domains. In these cases, it is highly preferable that the 1D domains do not control the timestep, as 99% of the computational effort is usually in solving the 2D domains.
As of Build 2005-05-AN, different timesteps can be specified for the 1D and 2D domains, offering much greater flexibility in setting timesteps and model resolutions.
Two options exist for specifying eddy viscosity for the 2D domains to approximate the effect of small-scale motions that cannot be modelled directly. Use the Viscosity Formulation and Viscosity Coefficient commands to set the formulation and coefficient.
The first method (Viscosity Formulation == CONSTANT) is to supply a constant value, E, which is used throughout the model. This is generally satisfactory when the cell size is much greater than the depth or when other terms are dominant (eg. high bed resistance).
The second method (Viscosity Formulation == SMAGORINSKY) is an approximation to the Smagorinsky formulation. This formulation is preferred where the cell size is similar or less than the depth.
Testing by Barton 2001 indicates that 2D schemes using very fine elements (less than 2m) may have difficulty predicting correct flow behaviour. Results from models with less than 2m cell size should be treated with caution, particularly if the depths are greater than the cell size and/or the friction forces are low (ie. low Manning’s n).
Dostları ilə paylaş: | 
