|
|
|
|
|
You are calculating the overlap between a cell and a sphere by using the bounding box of the cell. The bounding box is larger than the cell. So, sometimes this will return true when there is no an intersection; when the bounding box intersects the sphere but the cell does not. Yes? Can you improve the algorithm so that it doesn't produce false positives like this? The treeDataFace class also does not have this method implemented. Could you complete this implementation too? |
|
|
Indeed the algorithm might result in false positives for skewed and/or big cells (compared to the sphere being searched). I'll take a look how this can be implemented in an efficient manner. I can implement the same functionality for the "treeDataFace" class. I own the copyright associated with this potential contribution, but I have not signed any agreements. |
|
|
Will, this has now been implemented for both the treeDataCell and treeDataFace classes. I've attached a screenshot of a test case with this note. The red colored cells are those overlapped by the gray sphere. Similarly, the green small spheres represent the internal faces overlapped by that same sphere. testCase.png (132,209 bytes) |
|
|
Your algorithms still don't work in a variety of edge cases. See images. Both cell and face intersections are being missed. I don't think this is possible as a contribution. If you need this functionality implemented then please contact CFD Direct for support and/or development services. cells.png (27,251 bytes) faces.png (115,432 bytes) |
- Anonymous
