<html><head><style type='text/css'>p { margin: 0; }</style></head><body><div style='font-family: arial,helvetica,sans-serif; font-size: 12pt; color: #000000'>This is exactly what I mean, but I was talking about 3D.<br>Anyway, from what you're saying I cannot understand the use of GeometryGrid.<br>If I can move nodes using a deformation function and I can refine locally (in Dune style), the "moved" situation is quite common.<br>Moreover, in this case I'd sum over the 5 intersections of the right element asking for normals and intersections length, what is the relevance of being a cube for the element? Intersections are "cube".<br>I can agree if you speak about the volume of the right element.<br>Who's Z?? You meant Y?<br>Thanks,<br><br>Marco<br><br><br><div><span name="x"></span>--<br>-----------------------------------------------<br>Marco Cisternino, PhD<br>Software Developer<br>OPTIMAD Engineering s.r.l.<br>Via Giacinto Collegno 18<br>10143 Torino - Italy<br>www.optimad.it<br>marco.cisternino@optimad.it<br>+39 011 19719782<br>-----------------------------------------------<span name="x"></span><br></div><br><hr id="zwchr"><div style="color:#000;font-weight:normal;font-style:normal;text-decoration:none;font-family:Helvetica,Arial,sans-serif;font-size:12pt;"><b>Da: </b>"Andreas Dedner" <a.s.dedner@warwick.ac.uk><br><b>A: </b>"Marco Cisternino" <marco.cisternino@optimad.it><br><b>Cc: </b>dune@dune-project.org<br><b>Inviato: </b>Martedì, 24 febbraio 2015 18:00:30<br><b>Oggetto: </b>Re: [Dune] parametrized AMR and cell divergence<br><br>
  
    
  
  
    So you have the following situation<br>
      Original                                           Moved<br>
       X-----------A----------------X                
    X--------A------------------X           <br>
       |              |                    |                 
    |           \                      |<br>
       |              |                    |                 
    |            \                     |<br>
       |------------Y                    |                 
    |-----------Y                   |<br>
       |              |                    |                 
    |            /                     |<br>
       |              |                    |                  |         
    /                       |<br>
       X-----------B----------------X                
    X-------B-------------------X<br>
    Perhaps somebody else can add his thoughts but the right element
    after<br>
    movement is not a cube any more so it still will have the straight
    edge<br>
    between AB and your grid elements would not be non overlapping. That<br>
    is not allowed in DUNE. <br>
    <br>
    The right element does have 5 intersections and I guess two of those
    could be between AZ and ZB... But the main problem is the non DUNE
    style grid.<br>
    <br>
    Andreas<br>
    <br>
    <div class="moz-cite-prefix">On 24/02/15 16:35, Marco Cisternino
      wrote:<br>
    </div>
    <blockquote cite="mid:1998062480.4520831.1424795707773.JavaMail.root@optimad.it">
      
      <style>p { margin: 0; }</style>
      <div style="font-family: arial,helvetica,sans-serif; font-size:
        12pt; color: #000000">Absolutely non conforming.<br>
        I know I have a hanging node, but I cannot understand what you
        say: why do you talk about coarse and fine intersections??<br>
        If I move a hanging node I'm moving 4 intersections, isn't it?
        And my coarse element has 9 intersections, right?<br>
        Through its intersections the coarse element knows about the
        hanging node or not?<br>
        If not I cannot imagine the use of intersections.<br>
        <br>
        Thanks again.<br>
        Marco<br>
        <br>
        <hr id="zwchr">
        <div style="color:#000;font-weight:normal;font-style:normal;text-decoration:none;font-family:Helvetica,Arial,sans-serif;font-size:12pt;"><b>Da:
          </b>"Andreas Dedner" <a class="moz-txt-link-rfc2396E" href="mailto:a.s.dedner@warwick.ac.uk" target="_blank"><a.s.dedner@warwick.ac.uk></a><br>
          <b>A: </b>"Marco Cisternino"
          <a class="moz-txt-link-rfc2396E" href="mailto:marco.cisternino@optimad.it" target="_blank"><marco.cisternino@optimad.it></a>, <a class="moz-txt-link-abbreviated" href="mailto:dune@dune-project.org" target="_blank">dune@dune-project.org</a><br>
          <b>Inviato: </b>Martedì, 24 febbraio 2015 16:55:38<br>
          <b>Oggetto: </b>Re: [Dune] parametrized AMR and cell
          divergence<br>
          <br>
          Hi.<br>
          Conforming or non conforming? If non conforming you have a
          hanging<br>
          node and if you move coarse and fine intersections will not
          match (the<br>
          coarse grid element does not know about the hanging nor where
          it<br>
          is). Could that be your problem?<br>
          Best<br>
          Andreas<br>
          <br>
          <br>
          <div class="moz-cite-prefix">On 24/02/15 15:39, Marco
            Cisternino wrote:<br>
          </div>
          <blockquote cite="mid:1005665574.4510952.1424792351225.JavaMail.root@optimad.it">
            <style>p { margin: 0; }</style>
            <div style="font-family: arial,helvetica,sans-serif;
              font-size: 12pt; color: #000000">Hi Andreas, thank you for
              your reply.<br>
              I'm using GeometryGrid and ALU together. I provide ALU
              with a DiscreteDeformationFunction (my coarse element
              parametrization). Any time a new point appears in the mesh
              after a refinement I move that point using the
              DiscreteDeformationFunction.<br>
              The problem appears for local refinement. On the other
              hand, if I globally refine the mesh I have no problem
              about the divergence of the cells: it is zero everywhere.<br>
              However, if I refine only a bunch of cells, then the
              coarse ones, sharing intersection with the fine ones, have
              non-zero divergence.<br>
              I'm doing all this using the following methods:<br>
              intersection centerOuterUnitNormal<br>
              and codimension 1 volume method (choosing the element face
              corresponding to the intersection)<br>
              I hope it is clearer.<br>
              Thanks,<br>
              <br>
              Marco<br>
              <br>
              <br>
              <hr id="zwchr">
              <div style="color:#000;font-weight:normal;font-style:normal;text-decoration:none;font-family:Helvetica,Arial,sans-serif;font-size:12pt;"><b>Da:

                </b>"Andreas Dedner" <a class="moz-txt-link-rfc2396E" href="mailto:a.s.dedner@warwick.ac.uk" target="_blank"><a.s.dedner@warwick.ac.uk></a><br>
                <b>A: </b><a class="moz-txt-link-abbreviated" href="mailto:dune@dune-project.org" target="_blank">dune@dune-project.org</a><br>
                <b>Inviato: </b>Martedì, 24 febbraio 2015 12:14:33<br>
                <b>Oggetto: </b>Re: [Dune] parametrized AMR and cell
                divergence<br>
                <br>
                I'm not quite sure I understand the question. Are you
                seeing problems<br>
                when refining the grid using ALU's standard globalRefine
                method or do the problems<br>
                only appear when you use your local parametrization and
                after moving nodes around<br>
                (I'm not quite clear how you do that with ALU....).<br>
                Andreas<br>
                <br>
                <br>
                <div class="moz-cite-prefix">On 24/02/15 11:00, Marco
                  Cisternino wrote:<br>
                </div>
                <blockquote cite="mid:637362868.4442210.1424775619676.JavaMail.root@optimad.it">
                  <style>p { margin: 0; }</style>
                  <div style="font-family: arial,helvetica,sans-serif;
                    font-size: 12pt; color: #000000">Good morning,<br>
                    I would like to ask duners about deformed AMR.<br>
                    I'm working with body fitted mesh and ALUGrid.<br>
                  </div>
                </blockquote>
                <blockquote cite="mid:637362868.4442210.1424775619676.JavaMail.root@optimad.it">
                  <div style="font-family: arial,helvetica,sans-serif;
                    font-size: 12pt; color: #000000">In order to refine
                    my mesh I compute a local parametrization for every
                    coarse element, then I refine the mesh moving new
                    nodes using my parametrization.<br>
                    Therefore I compute the cell divergence summing over
                    the cell intersection<br>
                    normal . flux * area<br>
                    where intersection normals and areas are computed by
                    Dune methods centerUnitNormal and volume(this for a
                    face), while flux is (1,1,1) everywhere.<br>
                    I expect 0 over the entire mesh and this is true on
                    the coarse grid, but for a refined grid this is not
                    true for coarse cells sharing intersections with
                    finer cells.<br>
                    Does anyone see something like this? And is there a
                    cure? Am I doing something wrong?<br>
                    <br>
                    Thanks for any hint.<br>
                    Bests,<br>
                    <br>
                    Marco<br>
                    <br>
                  </div>
                  <br>
                  <fieldset class="mimeAttachmentHeader"></fieldset>
                  <br>
                  <pre>_______________________________________________
Dune mailing list
<a class="moz-txt-link-abbreviated" href="mailto:Dune@dune-project.org" target="_blank">Dune@dune-project.org</a>
<a class="moz-txt-link-freetext" href="http://lists.dune-project.org/mailman/listinfo/dune" target="_blank">http://lists.dune-project.org/mailman/listinfo/dune</a>
</pre>
                </blockquote>
                <br>
              </div>
              <br>
            </div>
          </blockquote>
          <br>
        </div>
        <br>
      </div>
    </blockquote>
    <br>
  

</div><br></div></body></html>