<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"><html xmlns="http://www.w3.org/1999/xhtml"><head>
<meta content="text/html; charset=UTF-8" http-equiv="Content-Type"/>
</head><body style="">
<div>
Dear gmsh/getdp community,
<br/>
<br/>The minimal example test.{geo,pro} demonstrates a problem I have with
<br/>the calculation of fluxes through partial surfaces.
<br/>
<br/>The Laplacian equation is solved for the potential v on the unit square
<br/>[0,1]^2 with natural boundary conditions at x=0 and x=1, with v=0 on
<br/>y=0, and v=1 on y=1.
<br/>
<br/>To get partial fluxes iA and iB through line segments x\in[0,split] and
<br/>x\in[split,1] at y=1 I am using two node groups and corresponding
<br/>global quantities.
<br/>
<br/>The problem is that I always get the superposition of the boundary
<br/>potentials at the intersection point of the intervals, i.e., v=2 at x=split, y=1.
<br/>
<br/>All attempts to exclude the intersection point from the physical
<br/>regions of the partial surfaces failed. (Some of these attempts are indicated by comments in
<br/>the .pro-file.)
<br/>
<br/>If I introduce a small gap between the line segments A and B I get
<br/>almost the expected fluxes.
<br/>
<br/>But the introduction of the small gap does not work for the real
<br/>application. The target is a 3D domain defined by a STL grid translated to the gmsh-geo format. The
<br/>boundary is defined by surfaces and the partial boundaries are defined
<br/>by separation curves. There, I have the same effect.
<br/>
<br/>With best regards,
<br/>Tobias Naehring
<br/>
<br/>
</div>
<div>
PS: Thank you for the tool.
</div>
<div>
</div>
</body></html>