We are using NX Generative Shape Design to approximate a precise defined CAD Surface to an tessellated one, e.g. from an .stl File.
It is actually the easiest and fastest way to convert tessellated data into precise (defined!) data. However, it also has some restrictions:
-Like tessellation only approximates triangles to describe curvatures, this solution only approximates a surface to triangles.
-the result will have no sharp edges.
-there are no simple parameters like lengths, angles or radii.
-depending on the complexity of the tessellated geometry and the desired accuracy of the result, the calculation can take a long time.
Let's start with the workflow.
Start NX, open a new part and import your tessellated geometry. (File -> Import -> STL)
For "Facet Body Output Type" select "Convergent". You can change this also afterwards using the command "Convert Facet Body".
Take also care of the resolution, if importing a geometry from STAR-CCM+ you should select STL File Units "m". if the geometry comes from another CAD program, it is usually "mm"
Now switch to the Surface Ribbon and start NX Realize Shape.
In case you don't have the "NX Realize Shape" Button in your menu, you can use the search field, or switch to "Content", "Role Advanced":
Now select the "Cage from Facet Body" Tool:
The tool is quite simple, just select your imported geometry. For the first try, you should leave the Average Face size at the default value or even increase it on slow machines.
Click "Show Result" and decide, if the outcome is already sufficient:
The following picture shows the tessellated input geometry:
This picture shows the Transformation Cage with default Average Face Size:
Inside the dialogue, click on OK and then on Finish on the Ribbon Bar to exit NX Realize Shape.
Now hide the Input geometry to review the Result:
As you can see, the edges of the Cylinder and the Cuboid are not looking really good, whereas the Sphere is close to what we want.
The comparison: The resulting Body has no sharp edges, where these were located before, the result is now dented. Round objects are approximated very well.
Increasing the Average Face Size is resulting in a more accurate result, but be careful: the time to calculate increases exponentially!
(in this case it took ~5min using Average Size of 10 instead of 48.5)
Conclusion:
In case you don't want to modify the geometry afterwards, you can live with the approximation of sharp-edged entities (tessellation is only able to approximate bend faces), the geometry itself is not too complex and you can live with a big file size, this is the right way.