The SGrid Cut Revisited

Laurent Deny and David DeBaun and Bruno Levy and Richard Cognot. ( 1997 )
in: Proc. $16^{th}$ Gocad Meeting, Dallas

Abstract

The Sgrid Cut is a complex algorithm aIlowing to disconnect ceIls of a grid in which a given surface pass through. This algorithm is based on the notion of Split Node, which is a node of the grid that can have several geometric locations, while it has single u,v,w indices within the grid. The CUITent version of the Sgrid Cut had to face several particular cases, occurring when several faults pass through the same node, or when a fault ends within the grid. The treatment of these cases lead to several improvements of the CUITent implementation. First, the algorithm is now based on the set of combinatorial operations one can apply to a SGrid, while the CUITent method is based on a data structure. This new approach aIlows to have a tight theoretical control on the method. This aIlowed to prove that the new implementation is able to deal with aIl the particular cases that may occur, instead of using an empiric approach. Moreover, the new objects developed are compatible by the new G-Maps based topological model of gOcad, thus enabling to represent complex relationships between different grids.

Download / Links

    BibTeX Reference

    @INPROCEEDINGS{Deny97b,
        author = { Deny, Laurent and DeBaun, David and Levy, Bruno and Cognot, Richard },
         title = { The SGrid Cut Revisited },
     booktitle = { Proc. $16^{th}$ Gocad Meeting, Dallas },
          year = { 1997 },
      abstract = { The Sgrid Cut is a complex algorithm aIlowing to disconnect ceIls of a grid in which a
    given surface pass through. This algorithm is based on the notion of Split Node, which is
    a node of the grid that can have several geometric locations, while it has single u,v,w
    indices within the grid. The CUITent version of the Sgrid Cut had to face several particular
    cases, occurring when several faults pass through the same node, or when a fault ends
    within the grid. The treatment of these cases lead to several improvements of the CUITent
    implementation. First, the algorithm is now based on the set of combinatorial operations
    one can apply to a SGrid, while the CUITent method is based on a data structure. This new
    approach aIlows to have a tight theoretical control on the method. This aIlowed to prove
    that the new implementation is able to deal with aIl the particular cases that may occur,
    instead of using an empiric approach. Moreover, the new objects developed are
    compatible by the new G-Maps based topological model of gOcad, thus enabling to
    represent complex relationships between different grids. }
    }