| Author | Message | ||
     Clemont Fourcade (clemont) New member Username: clemont Post Number: 1 Registered: 04-2007 |
Hello, I am trying to model a pretty complex problem. The problem can be divided in two stages. In the first one, a time-dependent heat source defines the temperature in time in a given cell. In the second stage we calculate the viscous flow (which depends on the time-dependent temperature defined in the first stage) but only in a small region of the previous cell. I think that using export-import tools I could solve the problem. However, I was not able to find information about export-import when a problem is time-dependent. I do not know either if when the two stages of the problem have different cell size would raise an additional problem (let's say than we calculate the temperature stage in a cell of width=(-L, L), depth=(0, -d) and the second stage we want to solve the flow equations in a cell with width=(-L/5, L/5), depth=(0, -d/5)). I have only been able to write the script for a simpler problem, when all the equations (temperature and viscosity flow) are calculated in the same stage, using hence the same cell size. However, this implies important numerical inaccuracies for the calculations, due to exponential relationship between the viscosity and temperature. I would be very happy to get any comment or suggestion about the problem. Thank you very much Clem | ||
| Robert G. Nelson (rgnelson) Moderator Username: rgnelson Post Number: 829 Registered: 06-2003 |
This is a problem. At present, there is no mechanism for reading in time-dependent transfer data and tracking the interpolation through the data as time progresses. The only way I can see to do this is to solve the temperature and flow simultaneously, defining a trivial flow equation in the outer region that generates the proper boundary conditions for the smaller region. There is an inefficiency incurred in carrying the flow variables in the region where they are not needed, but I don't see any way around it. |