Re: [deal.II] Re: variable time step in step44

2016-09-03 Thread Jean-Paul Pelteret
Dear Anup, The simplest explanation that I can give is that, in its original form, the implementation of the Time class in step-44 is not suitable for your needs and you need to adjust it accordingly. If you need to change that value of delta_t, then you should add some function that does it fo

Re: [deal.II] Re: variable time step in step44

2016-09-03 Thread Anup Basak
Dear Jean-Paul, Thank you for your reply. I, in fact, have done that way only. I have defined the initial time step through .prm file. Now my question is that after I update my time step in the run() function how can I pass this new time step, say 'delta_t' to the class 'Time' to obtain other par

[deal.II] Re: variable time step in step44

2016-08-31 Thread ha . badnava
Dear Anup, You can reduce current time increment by a “time scaling factor” (I am using 10 or 4 as time scaling factor) . However, it is better to define a “minimum time increment size” to stop simulation when the current time increment becomes less than it. In addition, when the current incre

[deal.II] Re: variable time step in step44

2016-08-30 Thread Jean-Paul Pelteret
Dear Anup, Why don't you reinterpret the parameter passed as "delta_t" to be the initial timestep size "delta_t_0" and then adjust the current timestep size "delta_t" however you want? How you do this adjustment algorithm-wise is, of course, entirely up to you. Regards, J-P On Tuesday, Augus