|
[Sponsors] |
February 20, 2019, 03:39 |
Speed of solver determined by what
|
#1 |
New Member
Join Date: Feb 2019
Posts: 15
Rep Power: 7 |
Hi all, I run interFoam for simulating water running through a channel. Sometimes I get a solution computed in 6 sec. By e.g. changing the amount of input water to lower amounts, the solver does not get even a first time step calculated after an hour so I kill the job. The same problem occurs if I refine the cells to half size. The solver does not get anywhere and keeps on computing. How can one avoid this, predict this or check this from the log to find out what is wrong? Thx
|
|
February 20, 2019, 05:01 |
|
#2 | ||
Senior Member
Robert
Join Date: May 2015
Location: Bremen, GER
Posts: 292
Rep Power: 12 |
Quote:
i haven't worked with interFoam so i cannot comment on the amout of water, but the issue with the cells will be the same for all solvers. Refine your cells by one level and you increase the number of cells by a factor of 8. Next refinement layer, another factor 8 increase and you end up with 64 cells instead of one. Of course, this increases the resolution of your model, but it also increase your runtime (and by the same factor i'd guess). Quote:
- Before creating your mesh, think about what you want to accomplish: Here, you are dealing with laminar flow through a pipe, so i might guess you want to sample a velocity profile. ~ 20 cells in width and height would be sufficient for that job. Dont create to many cells. - Grab the fvSolution and fvSchemes files from an existing OF case similar to yours. Dont meddle with these if do not know what you are doing. - Are you running a laminar or a transient case? IF laminar: Check your residual control settings. For a first test a value of 1e-4 would be sufficient. IF transient: Check your Courant number. You want to be below 1.0 but 0.99 is fine. You don't need to run your simulation with a Courant number off 0.001. 2. How can you predict this?: Keep in mind, more cells = more runtime. Co << 1 and you can increase your timestep. 3. What can you read from the log?: The courant number. Number of iterations. Residual values. The latter two should be decreasing, or at least, no increase over time. Doesn't tell you much in the way of what is wrong, more that something is amiss.
__________________
If you liked my answer to your question, please consider leaving a "Like" in return |
|||
Tags |
solver speed |
Thread Tools | Search this Thread |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
icoFoam Courant number growth problem | tdof | OpenFOAM Running, Solving & CFD | 12 | May 24, 2023 15:00 |
different results between serial solver and parallel solver | wlt_1985 | FLUENT | 11 | October 12, 2018 09:23 |
Can't get data from OpenFoam to external solver using externalCoupled | perry | OpenFOAM Running, Solving & CFD | 4 | May 26, 2014 09:09 |
OpenCL linear solver for OpenFoam 1.7 (alpha) will come out very soon | qinmaple | OpenFOAM Announcements from Other Sources | 4 | August 10, 2012 12:00 |
Working directory via command line | Luiz | CFX | 4 | March 6, 2011 21:02 |