Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.
Dividing Thermal Stresses problem into 2 study steps?
Posted 23 mars 2020, 12:47 UTC−4 Structural Mechanics Version 5.5 2 Replies
Please login with a confirmed email address before reporting spam
Would there be any downside in dividing a stationary thermal stresses problem into 2 study steps, where you first solve for only the heat transfer in step 1 and then solve only for solid mechanics in step 2 (obviously using the results from step 1)?
Please login with a confirmed email address before reporting spam
Hello Andreas,
If the two physics are only one-way coupled, then I cannot think of any downside, and of course there in the upside that your peak memory usage will be lower when you split the problem (I assume that's why you're thinking of splitting the problem, did I guess that right?).
On the other hand, if the two physics are two-way coupled in any way (such as if the heat sources or any other aspect of the thermal problem are deformation dependent), then you can't split the problem.
In other words: If you could theoretically split the problem when solving the same equations with pen and paper, then you can do it in COMSOL too. Best,
Jeff
-------------------Jeff Hiller
Please login with a confirmed email address before reporting spam
Actually, you can expect a speed-up by at least a factor of 3 if you split it. Further advantages:
- If one of the problems is nonlinear, you will not need to perform iterations for both
- When the temperature problem is time dependent, you can usually still solve the stress problem as a sequence of stationary 'snapshots'
Henrik Sönnerlind
COMSOL