[kmefun]

Hi Foamer,

I am implementing a marangoniBC follow by this thread
http://www.cfd-online.com/Forums/ope...angoni-bc.html

The velocity field result shows resonable from paraFoam.
U.jpg
However, the velocity patch field in each runTime directory show unrealistic values, like

right
{
type marangoni;
marangonicoeff -0.013;
value nonuniform List<vector>
20
(
(0 0 0)
(0 0 0)
(0 0 0)
(0 0 0)
(0 0 0)
(0 0 0)
(0 0 0)
(0 0 0)
(0 0 0)
(0 0 0)
(0 0 0)
(0 0 0)
(0 0 0)
(0 0 0)
(0 0 0)
(0 0 0)
(0 0 0)
(0 0 0)
(0 0 0)
(0 -0.00325 0)
)
;
}

When I run openFOAM in parallel and then reconstruct processors. The velocity patch field then can be got resonable distribution values, like

right
{
type marangoni;
marangonicoeff -0.013;
value nonuniform List<vector>
20
(
(0 -9.82226e-06 0)
(0 -4.04154e-05 0)
(0 -9.46266e-05 0)
(0 -0.000163121 0)
(0 -0.000241381 0)
(0 -0.000327055 0)
(0 -0.000417371 0)
(0 -0.00050884 0)
(0 -0.000597326 0)
(0 -0.000678225 0)
(0 -0.000746756 0)
(0 -0.000798225 0)
(0 -0.000828147 0)
(0 -0.000832275 0)
(0 -0.00080656 0)
(0 -0.000746588 0)
(0 -0.000645329 0)
(0 -0.000493089 0)
(0 -0.00031147 0)
(0 -0.000192022 0)
)
;
}
My question is no matter runing in single or parallel, the writing out value should be the same, right? The attachment are the boundary Patch I implemented.marangoniFvPatchVectorField.C

marangoniFvPatchVectorField.H
Could anybody give me some clue about it?

Thank you!!
I use OF 2.3 and the application is icoFoam with adding temperature equation in it.