Page 1 of 1

Maximu value for dt

PostPosted: Wed May 16, 2018 3:38 pm
by iyaniwura
Dear Moderator,

I'm solving a coupled PDE-ODE model using FlexPDE 6 and observe something that looks like numerical instability around time = 100. I observed that around this time, the dt is large, it is possible for me to specify a maximum value for dt. Maybe I can prevent this instability by keeping the dt as low as possible.

I have a copy of my code incase you will like to see it.

Thanks for your attention.

Sarafa.

Re: Maximu value for dt

PostPosted: Thu May 17, 2018 2:08 pm
by moderator
I ran this script on version 6.51 with the ERRLIM control removed (to let it run faster).
The oscillatory behavior is established from the start, and merely gets more extreme after time 90.
The timestep I observed is about 0.05, and the oscillation period is about 10, so there are 200 timesteps per oscillation.
This is not a timestep size issue.
The oscillation is inherent in your equations.

Re: Maximu value for dt

PostPosted: Thu May 17, 2018 5:47 pm
by iyaniwura
Thanks for the prompt response.

I was also thinking that the value of dt should not determine the stability of the solution since FlexPDE 6 uses BDF. My next guess will be the position of the cells, actually, I just discovered that I made a mistake when computing the positions of the cells.

I will check the code again and let you know if I am able to resolve the issue.

Thanks once again.