Tricks of the Trade: Only solve the "Problem"
 Often, CFD is used to simulate an entire complicated geometry when the specific problem could be solved by simulating just the relevant part. A global result is always great to have but when the total size of a particular problem is limited by computing power or licensing a more accurate result may be achieved by simulating just the relevant parts of the geometry. This allows higher mesh density in the important areas and prevents errors generated in a less relevant regions from being propagated into the area of interest.
As a Naval Architect you may need to think about whether you need to have a full freesurface model to get the best design for a hydrodynamic or aero dynamic feature, like a rudder or pilot house, or if it is better to focus on the feature itself and not waste grid capturing effects on the surface that don’t effect the design.
Tricks of the trade: Start small!
 CFD has a bit of a reputation for eating up time and dollars. Sometimes it can take days to generate a solution and then, after all that time, you realise there was a mistake in the setup.
A good way around this is to start your simulation with a small number of large cells, that way it’ll run quickly and you can iron out any setup and physics errors without having to wait days for the computer. Then when you’re ready put in more cells and get an accurate answer!
This won’t avoid every problem as some issues, particularly with convergence, don’t show up until the grid is finer but it’ll get you off to a good start. There’s nothing worse than waiting overnight for some precious results before realising you forgot to scale the geometry!
|