<html>
  <head>
    <meta content="text/html; charset=utf-8" http-equiv="Content-Type">
  </head>
  <body bgcolor="#FFFFFF" text="#000000">
    Hi Jon,<br>
    <br>
    Thanks for the quick response. I too have a sudden increase/decrease
    in temperature and internal energy in a single cell (I assume that
    E_int --> T --> P). It is possible that there is a single
    origin to this problem. Let us notify one another about any progress
    on this matter, and anyone else from the community.<br>
    <br>
    Best wishes,<br>
    Seyit<br>
    <br>
    <br>
    <br>
    <div class="moz-cite-prefix">On 06/02/17 19:24, Slavin, Jonathan
      wrote:<br>
    </div>
    <blockquote
cite="mid:CACcRS=cmfbGynWCo+Fzog4HHnVuhgCCfWxELiQkhKvDVgRBEhg@mail.gmail.com"
      type="cite">
      <div dir="ltr">
        <div class="gmail_extra">
          <div class="gmail_default"
            style="font-family:arial,helvetica,sans-serif;font-size:small">​Hi
            Seyit,</div>
          <div class="gmail_default"
            style="font-family:arial,helvetica,sans-serif;font-size:small"><br>
          </div>
          <div class="gmail_default"
            style="font-family:arial,helvetica,sans-serif;font-size:small">I
            have had to put that problem aside for a while, though I
            hope to get back to it soon.  I have not solved the problem
            that leads to negative or zero timesteps.  What I found was
            that what leads to the problem is that suddenly in a single
            grid zone/pixel the thermal pressure jumps to a very high
            value, which thus leads to a very small timestep.  I plan to
            probe the problem further as soon as I can.</div>
          <div class="gmail_default"
            style="font-family:arial,helvetica,sans-serif;font-size:small"><br>
          </div>
          <div class="gmail_default"
            style="font-family:arial,helvetica,sans-serif;font-size:small">Regards,</div>
          <div class="gmail_default"
            style="font-family:arial,helvetica,sans-serif;font-size:small">Jon​</div>
          <br>
          <div class="gmail_quote">On Mon, Feb 6, 2017 at 1:00 PM, <span
              dir="ltr"><<a moz-do-not-send="true"
                href="mailto:flash-users-request@flash.uchicago.edu"
                target="_blank">flash-users-request@flash.uchicago.edu</a>></span>
            wrote:<br>
            <blockquote class="gmail_quote" style="margin:0 0 0
              .8ex;border-left:1px #ccc solid;padding-left:1ex">
              <div id=":1ew" class="a3s aXjCH m15a14952389b85ad">From: Seyit
                Hocuk <<a moz-do-not-send="true"
                  href="mailto:seyit@mpe.mpg.de">seyit@mpe.mpg.de</a>><br>
                To: "<a moz-do-not-send="true"
                  href="mailto:flash-users@flash.uchicago.edu">flash-users@flash.uchicago.edu</a>"
                <<a moz-do-not-send="true"
                  href="mailto:flash-users@flash.uchicago.edu">flash-users@flash.uchicago.edu</a>><br>
                Cc: <br>
                Date: Mon, 6 Feb 2017 10:56:22 +0100<br>
                Subject: [FLASH-USERS] MHD simulations Eint=0 problem<br>
                Hi community,<br>
                <br>
                Has anyone else encountered this problem before?<br>
                <br>
                Whenever I run MHD simulations with the USM solver
                (Flash4.4), I occasionally get Eint=0. This seems to
                occur only when I put a B-field > 0 (I often simulate
                mass-to-flux ratio of 2). The problem is bigger with
                higher magnetic field strength. I have played around
                with the "eintswitch" parameter, thereby reducing
                reducing the issue it as much as I can (eintswitch =<
                1e-4), but the problem still persists. A high
                eintswitch, on the other hand, thereby reverting to an
                independent internal energy calculation was giving me
                crazy values. It is as if the magnetic energy is not
                considered for the internal energy calculations when
                derived from the total energy (Eint = E_tot - Ekin -
                Emag), or something along those lines, but this is just
                guessing.<br>
                <br>
                Optimizing the parameters helped a little. See parameter
                choices below. My initial guess was that the solver has
                a hard time solving, because the problem is stiff on
                occasion. Playing with these options and relaxing the
                solver still did not completely remove this issue.<br>
                <br>
                I initially thought that this happens after new cells
                are created, however, the problem does still occur when
                lrefine_min=lrefine_max.<br>
                <br>
                The problem also seems to persist if:<br>
                ForceHydroLimit = .True. (but the B-field is not set to
                zero; in the tested case it is 10.0e-6)<br>
                So, forcing hydrolimit did not help! This while when the
                B = 0 (+Forcehydrolimit=true), the Eint does not become
                0, but very low instead (which is a different issue?).
                While with B=0, I also get a different refinement.
                Therefore, I am not entirely sure what forcehydrolimit
                does actually. Clearly, there is still impact by the
                B-field.<br>
                <br>
                Another funny thing is enableMaskedGCFill. I usually
                have<br>
                enableMaskedGCFill = .True.<br>
                but the module that made use of this is turned off.
                Setting this to true or false gives me somewhat
                different kind of errors. With .False. I also get
                negative values, i.e., WARNING after gc filling: min.
                unk(EINT_VAR)=-1487770975.0815<wbr>08<br>
                <br>
                The end results is sometimes (especially if I don't
                constrain by "small" variables, i.e., make the "small"
                values really small):<br>
                DRIVER_ABORT: Computed new time step smaller than
                dr_dtMinContinue!<br>
                similar to a problem mentioned recently "dt not positive
                in MHD simulations" by Jon Slavin. Was any solution
                fouond for that?<br>
                <br>
                What does help is:<br>
                cfl = crazy low, but normally lowering cfl only delays
                the issue.<br>
                <br>
                What doesn't help is:<br>
                Turning off gravity<br>
                Turning off particles<br>
                Turning off refinement<br>
                Turning of all my modules (Heating/Cooling/Chemistry/Col<wbr>mndensities...)<br>
                Changing boundaries from periodic/isolated to outflow<br>
                Viscosity<br>
                And many other things<br>
                <br>
                With kind regards,<br>
                Seyit<br>
                <br>
                <br>
                Normal conditions (but I tested them all):<br>
                mfieldstrength = 5.0e-6<br>
                killdivb = .True.<br>
                E_upwind = .True.<br>
                energyFix = .True.<br>
                RiemannSolver = "HYBRID"<br>
                slopeLimiter = "HYBRID"<br>
                order = 3<br>
                use_steepening = .False.<br>
                use_flattening = .True.<br>
                use_avisc = .True.<br>
                shockDetect = .True.<br>
                cfl = 0.6<br>
                <br>
                Libraries/Compilers:<br>
                hdf5-serial/1.8.15<br>
                ifort: intel/15.0<br>
                impi/5.0.3<br>
                x86_64 x86_64 x86_64 GNU/Linux<br>
                <br>
                <br>
                -- <br>
                Dr. Seyit Hocuk<br>
                Post-doctoral researcher<br>
                Center for Astrochemical Studies (CAS)<br>
                Max-Planck-Institute for Extraterrestrial physics (MPE)<br>
                Giessenbachstrasse 1, 85748 Garching bei Munchen<br>
                Phone: <a moz-do-not-send="true"
                  href="tel:%2B49-8930000-3012" value="+4989300003012"
                  target="_blank">+49-8930000-3012</a> | Fax: <a
                  moz-do-not-send="true" href="tel:%2B49-8930000-3569"
                  value="+4989300003569" target="_blank">+49-8930000-3569</a><br>
                Web  : <a moz-do-not-send="true"
                  href="http://www.mpe.mpg.de/%7Eseyit" rel="noreferrer"
                  target="_blank">www.mpe.mpg.de/~seyit</a><br>
                <br>
              </div>
            </blockquote>
          </div>
          <br>
          <br>
          <br clear="all">
          <div><br>
          </div>
          -- <br>
          <div class="gmail_signature" data-smartmail="gmail_signature">
            <div dir="ltr">
              <div>
                <div dir="ltr">________________________________________________________<br>
                  Jonathan D. Slavin                 Harvard-Smithsonian
                  CfA<br>
                  <a moz-do-not-send="true"
                    href="mailto:jslavin@cfa.harvard.edu"
                    target="_blank">jslavin@cfa.harvard.edu</a>       60
                  Garden Street, MS 83<br>
                  phone: (617) 496-7981       Cambridge, MA 02138-1516<br>
                  cell: (781) 363-0035             USA<br>
________________________________________________________<br>
                  <br>
                </div>
              </div>
            </div>
          </div>
        </div>
      </div>
    </blockquote>
    <br>
    <pre class="moz-signature" cols="72">-- 
Dr. Seyit Hocuk
Post-doctoral researcher
Center for Astrochemical Studies (CAS)
Max-Planck-Institute for Extraterrestrial physics (MPE)
Giessenbachstrasse 1, 85748 Garching bei Munchen
Phone: +49-8930000-3012 | Fax: +49-8930000-3569
Web  : <a class="moz-txt-link-abbreviated" href="http://www.mpe.mpg.de/~seyit">www.mpe.mpg.de/~seyit</a></pre>
  </body>
</html>