[FLASH-USERS] noisiness in run with high ekin/eint

[yueh wang]

Jon,
I forgot to say;
V=r/t, and Rho =  r^(-n)* t^(-3) for the power-law ejecta. You need to have a non-zero starting time.
Given the time, V is just proportional to r.
The built-in BC won't work for SNR simulations.
Hope this help.
- Yueh


Jon,

The free expansion ejecta must have a velocity profile V=R/T.
Your inner BC is NOT placed at zero, which is good. Inside the inner BC, i.e. the inn ner ghostzones, you still need to set up the velocity and density profile so that the flow satisfies the free expansion. If the free expansion is not satisfied, the flow will first clash at the inner boundary, and the numbers (pressure and density) may eventually run 'wild.'

1-D spherical POLAR symmetry: don't use theta near 0 (the pole). Try theta (th) at theta=pi/2 (equator) and a reasonably small delta theta (dth).

I think some trials and errors on the BC (and IC) would help in this cases.

-Yueh



________________________________
From: flash-users-bounces at flash.uchicago.edu <flash-users-bounces at flash.uchicago.edu> on behalf of Slavin, Jonathan <jslavin at cfa.harvard.edu>
Sent: Thursday, March 29, 2018 9:29 PM
To: flash-users
Subject: [FLASH-USERS] noisiness in run with high ekin/eint

Hi,

I've been doing runs in 1D spherical symmetry of supernova remnants starting with fast cool ejecta.  This is one of those situations in which the kinetic energy far exceeds the thermal energy (at least in the early stages).

I know about eintSwitch which is supposed to help with numerical issues that arise in such situations, but I've been finding that setting it lower actually results in less noisiness, though the noisiness that remains is, if nothing else, unsightly.  What happens is that the pressure develops fairly large fluctuations that seem to be purely numerical since there doesn't seem to be any other source for them.  The pressure is small in those regions, relative to the kinetic energy, but still the noise seems to affect the density.

So I'm wondering if others have found parameter settings or approaches that work well in this sort of situation.  If so, I'd like to hear about them.  I should mention that I've been using split PPM.  I did try to use unsplit hydro but it failed with the pressure dropping to the smallp value very quickly.  I may try to investigate that some more.

Regards,
Jon

--
________________________________________________________
Jonathan D. Slavin                 Harvard-Smithsonian CfA
jslavin at cfa.harvard.edu<mailto:jslavin at cfa.harvard.edu>       60 Garden Street, MS 83
phone: (617) 496-7981       Cambridge, MA 02138-1516
cell: (781) 363-0035             USA
________________________________________________________

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://flash.rochester.edu/pipermail/flash-users/attachments/20180331/d09f81b6/attachment.htm>