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.

Solution Changes with time

Please login with a confirmed email address before reporting spam

Hello every body,

I have a problem and hope somebody can help. I compute a laminar two phase flow setting for for example 0 to 100 seconds and the solution is something. then when I compute the same problem with the same setting for for example 0 to 400 seconds, the solution is totally different; even the solution of 0 to 100 seconds portion of the whole solution is not the same as the case when I compute for 0 to 100 seconds. why should this happen??? note that the flow is transient but there is no transient boundary condition.
Thank You All

4 Replies Last Post 31.07.2015, 09:25 GMT-4
Josh Thomas Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 9 years ago 30.07.2015, 10:20 GMT-4
Hamid-

The initial time step size is dependent on the total time request (in your case, 400 seconds or 100 seconds). The initial step size can have an effect on the accuracy of the returned results.

For a one-to-one comparison and to confirm this is the issue you can try to fix the initial step size to make it the same for both cases (override the default initial time step size).

If indeed this is the issue, you may try selecting Time Stepping method as Strict rather than Free and selecting a smaller time step in your time range request. This will force the solver to be more accurate. You can also try tightening the absolute and relative tolerances for your problem. This should cause the 2 cases you described to converge to the same answer.

--
Best regards,
Josh Thomas
AltaSim Technologies
Hamid- The initial time step size is dependent on the total time request (in your case, 400 seconds or 100 seconds). The initial step size can have an effect on the accuracy of the returned results. For a one-to-one comparison and to confirm this is the issue you can try to fix the initial step size to make it the same for both cases (override the default initial time step size). If indeed this is the issue, you may try selecting Time Stepping method as Strict rather than Free and selecting a smaller time step in your time range request. This will force the solver to be more accurate. You can also try tightening the absolute and relative tolerances for your problem. This should cause the 2 cases you described to converge to the same answer. -- Best regards, Josh Thomas AltaSim Technologies

Josh Thomas Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 9 years ago 30.07.2015, 10:22 GMT-4
You may also check that the initial values for dependent variables is the same in both cases. Sometimes if you re-solve a previously solved model the legacy solution is defaulted to the initial values for the subsequent solution.

--
Best regards,
Josh Thomas
AltaSim Technologies
You may also check that the initial values for dependent variables is the same in both cases. Sometimes if you re-solve a previously solved model the legacy solution is defaulted to the initial values for the subsequent solution. -- Best regards, Josh Thomas AltaSim Technologies

Please login with a confirmed email address before reporting spam

Posted: 9 years ago 31.07.2015, 04:01 GMT-4
Dear Josh
Hi,

in fact, the range of my initial problem is (0,0.05,100) after that when I try to see what happens after 100 seconds, I compute the same problem and letting everything unchanged for (0,0.05,400) the solution is totally different. consider a channel for example when I compute the problem for (0,0.05,100) the direction of the flow is from left to right and when I compute the same problem for (0,0.05,400) the flow direction is from right to left. I changed the time stepping method from free to strict but nothing changed. what is your advice?

Thank you very much
Dear Josh Hi, in fact, the range of my initial problem is (0,0.05,100) after that when I try to see what happens after 100 seconds, I compute the same problem and letting everything unchanged for (0,0.05,400) the solution is totally different. consider a channel for example when I compute the problem for (0,0.05,100) the direction of the flow is from left to right and when I compute the same problem for (0,0.05,400) the flow direction is from right to left. I changed the time stepping method from free to strict but nothing changed. what is your advice? Thank you very much

Josh Thomas Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 9 years ago 31.07.2015, 09:25 GMT-4
Hamid-

I would check the other suggestions I gave in my previous posts specifically:

1) Fix the initial time step for both cases. Compare results. If changing to "strict" doesn't affect answer you may also need to decrease the time step request (currently 0.05).
2) Have you confirmed that the initial values of variables solved for is the same in both cases? You can check this by using the "Get initial values for step" option instead of "compute".

You can also post your model if you want people to get a better look at it.

--
Best regards,
Josh Thomas
AltaSim Technologies
Hamid- I would check the other suggestions I gave in my previous posts specifically: 1) Fix the initial time step for both cases. Compare results. If changing to "strict" doesn't affect answer you may also need to decrease the time step request (currently 0.05). 2) Have you confirmed that the initial values of variables solved for is the same in both cases? You can check this by using the "Get initial values for step" option instead of "compute". You can also post your model if you want people to get a better look at it. -- Best regards, Josh Thomas AltaSim Technologies

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.