SMS:Q&A CMS-Flow: Difference between revisions
No edit summary |
No edit summary |
||
| Line 4: | Line 4: | ||
'''Q:''' For a given CMS run, how much longer (by percentage) will that same run be with the sediment transport calculation enabled? | '''Q:''' For a given CMS run, how much longer (by percentage) will that same run be with the sediment transport calculation enabled? | ||
'''A:''' The increase in run time will depend on the sediment transport timing parameters that are set by the user and also which sediment transport option is selected. For total load formulations ([[SMS:Lund_Cirp_and_Watanabe_Formula|Watanabe and Lund_CIRP]]), I generally use a sediment transport time step in the range of 60 to 120 sec and a morphology change time step of 0.25 ht to 1 hr, depending on the application. If the hydrodynamic time step is in the range of say 0.5 to 10 sec (typical for most applications), then the sediment transport calculations are conducted relatively infrequently and add minimal computational burden (you may not even notice the difference). If the advection-diffusion equation is selected for sediment transport, the sediment transport time step is constrained by a Courant-type condition so you have to use a much smaller sediment transport time step that with the total load formulations. You could start out with a sediment transport time step of say 5 times the hydrodynamic time step. If that sediment transport time step is too large, it will be modified internally to maintain stability. The advection-diffusion equation is much more computationally intensive than the total load formulations and takes a lot more time to run. In cases that I have run, using the A-D equation has doubled the computation time. The morphology change time step is not dependent on selection of sediment transport formulation. | '''A:''' The increase in run time will depend on the sediment transport timing parameters that are set by the user and also which sediment transport option is selected. For total load formulations ([[SMS:Lund_Cirp_and_Watanabe_Formula|Watanabe and Lund_CIRP]]), I generally use a sediment transport time step in the range of 60 to 120 sec and a morphology change time step of 0.25 ht to 1 hr, depending on the application. If the hydrodynamic time step is in the range of say 0.5 to 10 sec (typical for most applications), then the sediment transport calculations are conducted relatively infrequently and add minimal computational burden (you may not even notice the difference). If the advection-diffusion equation is selected for sediment transport, the sediment transport time step is constrained by a Courant-type condition so you have to use a much smaller sediment transport time step that with the total load formulations. You could start out with a sediment transport time step of say 5 times the hydrodynamic time step. If that sediment transport time step is too large, it will be modified internally to maintain stability. The advection-diffusion equation is much more computationally intensive than the total load formulations and takes a lot more time to run. In cases that I have run, using the A-D equation has doubled the computation time. The morphology change time step is not dependent on selection of sediment transport formulation. | ||
| Line 11: | Line 9: | ||
'''Q:''' According to the user manuel of CMS, when a flow rate boundary condition is given to a cell string the momentum equation is not solved and the flow component normal to the face is prescribed. Is the continuity equation solved? | '''Q:''' According to the user manuel of CMS, when a flow rate boundary condition is given to a cell string the momentum equation is not solved and the flow component normal to the face is prescribed. Is the continuity equation solved? | ||
'''A:''' The continuity equation is not being solved since you specify the flow rate. The momentum and continuity equations are used to find the flow rate. | '''A:''' The continuity equation is not being solved since you specify the flow rate. The momentum and continuity equations are used to find the flow rate. | ||
Revision as of 19:38, 5 November 2012
Q: For a given CMS run, how much longer (by percentage) will that same run be with the sediment transport calculation enabled?
A: The increase in run time will depend on the sediment transport timing parameters that are set by the user and also which sediment transport option is selected. For total load formulations (Watanabe and Lund_CIRP), I generally use a sediment transport time step in the range of 60 to 120 sec and a morphology change time step of 0.25 ht to 1 hr, depending on the application. If the hydrodynamic time step is in the range of say 0.5 to 10 sec (typical for most applications), then the sediment transport calculations are conducted relatively infrequently and add minimal computational burden (you may not even notice the difference). If the advection-diffusion equation is selected for sediment transport, the sediment transport time step is constrained by a Courant-type condition so you have to use a much smaller sediment transport time step that with the total load formulations. You could start out with a sediment transport time step of say 5 times the hydrodynamic time step. If that sediment transport time step is too large, it will be modified internally to maintain stability. The advection-diffusion equation is much more computationally intensive than the total load formulations and takes a lot more time to run. In cases that I have run, using the A-D equation has doubled the computation time. The morphology change time step is not dependent on selection of sediment transport formulation.
Q: According to the user manuel of CMS, when a flow rate boundary condition is given to a cell string the momentum equation is not solved and the flow component normal to the face is prescribed. Is the continuity equation solved?
A: The continuity equation is not being solved since you specify the flow rate. The momentum and continuity equations are used to find the flow rate.
