Problem setting up the Fluid-transfer BC

Hi everybody,
I am setting up the Fluid-transfer BC along rivers..
My supermesh has been refined over a shapefile representing my rivers.
Then I imported an ASCII file with some points close to the rivers.
I linked the head reference from the ASCII file to the Fluid-transfer BC. Then I chose as Data Regionalization Method, 1D linear interpolation along lines and as Selection Map the shapefile with the rivers. I left everything else as default.
I selected all the points along the rivers shapefiles and I applied the parameter association.

But in some parts I got too many lines connecting different nodes.
You can see in the picture below where in yellow are the selected points along the river shapefiles where I applied the boundary conditions.
In red the points from the ASCII file I used for linking the parameters. Close to the intersections of two rivers I got a net of green lines.
I guess this is physically not correct. How can I work around this problem?

[url=http://imageshack.us/f/577/feflowsnapshot.png/]http://imageshack.us/f/577/feflowsnapshot.png/[/url]

Thank you very much.

This difficulty is caused by the fact that FEFLOW does not allow selecting faces for BCs, but chooses the faces from selected nodes. In such a case we typically refine the mesh a little bit and and/or move some nodes (Moving nodes will have to be done in the Classic interface at the moment), until we get two elements on top connected to the node where the rivers meet. Both of these elements will have one node without the BC, so the top face will not be covered between the rivers.

Thanks Peter.

Is it possible to overcome the problem by using another interpolation method?

No, the behaviour does not depend on the interpolation method. As you can see, the nodes are perfectly selected. This is just a question of how FEFLOW interprets element faces with fluid-transfer BCs set to all nodes - and for FEFLOW then the face carries the boundary condition.