<div dir="ltr">Hi Riccardo :<div><br></div><div> FLASH uses synchronous timestepping, so the the dt is the same across all blocks. This simplifies the evolution algorithm considerably. </div><div><br></div><div> Your simulation should be able to support a larger CFL number, even in 2D axisymmetry. I just checked, and in Ji et al, 2013 (<a href="http://adsabs.harvard.edu/abs/2013ApJ...773..136J" target="_blank">http://adsabs.harvard.edu/<wbr>abs/2013ApJ...773..136J</a>), we used a CFL of 0.4 in 2D axisymmetry with unsplit MHD, though this was on a uniform mesh. The fact that the crash is occurring early on does suggest an initialization issue, or possibly a boundary issue in your run. I'd suggest outputting a plotfile after the first timestep, and inspecting both it and the zeroeth timestep plotfile very carefully. Is anything unusual occurring with the velocity and/or sound speed? If so, where does the issue arise? </div><div> </div><div> Best wishes,</div><div><br></div><div> Bob</div><div><br></div><div><br></div><div> </div></div><div class="gmail_extra"><br><div class="gmail_quote">On Fri, Apr 28, 2017 at 11:07 AM, Riccardo Ciolfi <span dir="ltr"><<a href="mailto:riccardo.ciolfi@oapd.inaf.it" target="_blank">riccardo.ciolfi@oapd.inaf.it</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><br>
Dear Robert and Klaus,<br>
<br>
thank you very much for you reply.<br>
<br>
First, I would like to ask a clarification: does "global" delta_t mean that it is set the same in the entire block (but can still vary in different blocks) or it is the same in the entire computational domain of the simulation (i.e. in all the blocks)?<br>
<br>
To answer your question, my setup is purely hydro (an initial ball of matter in expansion, with a max expansion speed of 1e-4 c).<br>
In 1D it is a spherically simmetric expansion and cfl=0.8 is fine. In 2D I use axisymmetry and cylindrical coordinates with reflective bounday conditions on R_cyl=0 and outflow-type on R_cyl_max, z_min, and z_max.<br>
Could it be that the cfl is required very low close to the axis, thus compromising the choice of delta_t for the entire domain?<br>
<br>
The failure in 2D with any cfl larger than 0.1 is essentially at the beginning of the simulation.<br>
<br>
Thanks a lot!<br>
<br>
Best,<br>
Riccardo<br>
<br>
<br>
<br>
----- Messaggio originale -----<br>
Da: "Robert Fisher" <<a href="mailto:rfisher1@umassd.edu">rfisher1@umassd.edu</a>><br>
A: "Klaus Weide" <<a href="mailto:klaus@flash.uchicago.edu">klaus@flash.uchicago.edu</a>><br>
Cc: "Riccardo Ciolfi" <<a href="mailto:riccardo.ciolfi@oapd.inaf.it">riccardo.ciolfi@oapd.inaf.it</a>><wbr>, <a href="mailto:flash-users@flash.uchicago.edu">flash-users@flash.uchicago.edu</a><br>
Inviato: Venerdì, 28 aprile 2017 16:04:32<br>
Oggetto: Re: [FLASH-USERS] courant factor<br>
<div class="HOEnZb"><div class="h5"><br>
Dear Riccardo :<br>
<br>
As Klaus indicates, a CFL number of 0.8 is generally sufficient, and 0.5<br>
is conservative. Requiring 0.1 may indicate some other issues in your<br>
initial condition and/or boundary conditions, which may be physical, but<br>
could also be a bug.<br>
<br>
Can you also tell us when the crash is occurring? Shortly after<br>
initialization or well into the run? This is an important clue as to what<br>
is happening.<br>
<br>
It should be emphasized that the CFL condition is a necessary but not<br>
sufficient condition for numerical stability of a hydrodynamics code. This<br>
is because the condition itself is derived from a linearized analysis and<br>
can break down during strongly nonlinear evolution. For example, a strong<br>
acceleration or rapid cooling can drive the system unstable during a single<br>
timestep. For this reason, you will find problems with strong gradients in<br>
the initial condition, like shock tubes, often utilize a reduced initial<br>
timestep, and then "relax" to CFL. A reduced timestep can be set with the<br>
FLASH parameter dtinit.<br>
<br>
Best wishes,<br>
<br>
Bob<br>
<br>
On Thu, Apr 27, 2017 at 8:35 PM, Klaus Weide <<a href="mailto:klaus@flash.uchicago.edu">klaus@flash.uchicago.edu</a>><br>
wrote:<br>
<br>
> On Thu, 27 Apr 2017, Riccardo Ciolfi wrote:<br>
><br>
> > I am trying to understand the definition of the variable cfl. Does it<br>
> > correspond to the courant factor Delta_t/Delta_x * v_max?<br>
><br>
> The timestep Delta_t is chosen such that<br>
><br>
> Delta_t/Delta_x * v_max <= cfl<br>
><br>
> should be true in *each* cell (in 1D).<br>
><br>
> > Is this the same at all levels, so that Delta_t is reduced by a factor<br>
> > of 2 together with Delta_x when moving to the next refinement level?<br>
><br>
> There is only one global Delta_t. Delta_x depends on the block.<br>
><br>
> You can think of it as computing a local Delta_t for each cell, and the<br>
> global Delta_t is then the min of them.<br>
><br>
> > Is there a simple way to enforce it?<br>
><br>
> FLASH already does this for you.<br>
><br>
> > I am using FLASH4.4 and for 1D hydro simulations I adopted the suggested<br>
> > clf=0.8 without any problem. When I move to 2D I cannot run unless I put<br>
> > cfl as low as 0.1. Is this expected?<br>
><br>
> No, you should be able to uses larger values of cfl - as long as your<br>
> setup is purely a hydro problem.<br>
><br>
> > For coordinates with<br>
> > Delta_x=Delta_y, shouldn't be sufficient to reduce the courant factor by<br>
> > a factor of sqrt(2)?<br>
><br>
> See Hydro_computeDt for how exactly the computation is done for 2D and 3D.<br>
> But generally, a value of 0.8 or so should be good even for 2D.<br>
><br>
> Are you using curvilinear coordinates?<br>
> Are you using additional physics that change the solution?<br>
> Do you have unusual or extreme initial or boundary conditions?<br>
><br>
><br>
> Klaus<br>
><br>
<br>
<br>
<br>
--<br>
Dr. Robert Fisher<br>
Associate Professor / Graduate Program Director<br>
University of Massachusetts/Dartmouth<br>
Department of Physics<br>
285 Old Westport Road<br>
North Dartmouth, Massachusetts 02747<br>
<a href="mailto:robert.fisher@umassd.edu">robert.fisher@umassd.edu</a><br>
<a href="http://www.novastella.org" rel="noreferrer" target="_blank">http://www.novastella.org</a><br>
</div></div></blockquote></div><br><br clear="all"><div><br></div>-- <br><div class="gmail_signature" data-smartmail="gmail_signature"><div dir="ltr">Dr. Robert Fisher<br>Associate Professor / Graduate Program Director<br>University of Massachusetts/Dartmouth<br>Department of Physics<br>285 Old Westport Road<br>North Dartmouth, Massachusetts 02747<br><a href="mailto:robert.fisher@umassd.edu" target="_blank">robert.fisher@umassd.edu</a><br><a href="http://www.novastella.org" target="_blank">http://www.novastella.org</a></div></div>
</div>