Dear Bruno,
On Thursday, September 15, 2016 at 5:37:28 PM UTC+5:30, Bruno Turcksin wrote:: > > 1. What is wrong with the embedded method? Does the problem lie in my >> understanding? >> > It's hard to say what is wrong from here. Is coarsen_param greater than > one? Have you reached max_delta? Finally, just to be clear, the embedded > methods will use the time step that you supply if the error is less than > than refine_tol. However, if the time step can be coarsen, you can use > get_status().delta_t_guess to get a new time step that you can then use, > the next time you call evolve_one_time_step. > I'm indeed resetting my time step using get_status().delta_t_guess for the next step. The step has not reached max_delta, coarsen_param is set to 1.2. Any other reasons this can happen? > 2. Are embedded methods(with small time steps) a good choice for stiff >> problems or should I switch to an implicit method? >> > You probably will have to use an implicit method. The problem with > embedded methods for stiff problems is that the time steps will probably > become very small and the simulation will take for ever. > The reason I was trying to avoid an implicit method is that it takes more time per time step. I had earlier implemented my own version of the implicit Euler(by formulating residuals and performing Newton iterations), which took really long to run. The explicit methods run faster because they don't need Newton iterations at every step. However, I guess I'll need to stick with implicit methods if this doesn't work out. Thanks Vaibhav -- The deal.II project is located at http://www.dealii.org/ For mailing list/forum options, see https://groups.google.com/d/forum/dealii?hl=en --- You received this message because you are subscribed to the Google Groups "deal.II User Group" group. To unsubscribe from this group and stop receiving emails from it, send an email to dealii+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
