I am using Gridgen 15.00 to generate a grid from a ct images geometry (pulmonnary artery). I would like to extend the main branch with a cylinder. to do so, i have to connect a circle with the imported geometry outlet border. that is to say, to create a ruled surface between the circle and the outlet border. I split this problem in 2 halves, so i deal with a semi circle and a polylines in 3d (one half of the outlet border). The thing is, the control points of the polylines are not coplanar. when I try to create the ruled surface, there is no message of error, but the usual 2 segments connecting the starting and ending points of he 2 rail curves does not appear, and, more over, after clicking on "save ruled surface", the surface does not appears anywhere (in examine, delete, even in the connectors menu). I tried with a 7 control points polylines and a semi circle: if the 7 points are coplanar, everything is fine, but if they are not, i have the same problem. Is there a way to overcome it?

Hi jjc. First thing I'll recommend is that you upgrade to the latest version of Gridgen which is 15.13. If you're using V15.00, there have been many improvements since then. (You probably get tired of software vendors always telling you to upgrade to the latest!)

My first reaction to the problem you describe is that there's a tolerance issue here. Gridgen's default model size is 1000 units and if your CT data differs significantly (for example, if it's only 1 unit long) that could cause a problem.

If it's possible, can you email me your geometry (DBA file)? That would help me see exactly what's going on.

Also, you might try creating a Sweep Surface instead of a Ruled surface. As you know, the ruled surface works by interpolating between two opposite edges. The sweep surface is made by sweeping a profile along a line. It may avoid the problem you're seeing.

Best regards.

Thanks for your help.

I have found a solution of this problem by changing the tolerance value. The thing is, I don't know if this may have a significant impact on my further fluent computation or not. I will have a look at the sweep surface you've mentionned. By the way, do you know if an algorithm to create the centerline of a tetrahedral mesh already exists? I am both trying to find a way to do it myself and seeking for a solution available on the internet. thank you for your precious help. Best regards.

jjc: I don't understand why this would have an effect on the computation in Fluent. The tolerance value in Gridgen simply assures that the geometric computations are performed correctly. When you export to Fluent, the mesh should be just fine.

No, I'm not aware of an algorithm for finding the centerline of a tet mesh.