8.19.1. Solution Technique for Non-Conformal Boundaries

When two sections are associated along a non-conformal boundary, a geometrical tolerance dictates the search for correspondences. For every "source" boundary degree of freedom (at every mesh node, midface node, and so on), a correspondence is found in terms of a linear combination of “target" boundary degrees of freedom.

So source and target boundaries do not have to be coincident in the element connectivity sense; continuity is enforced by constraining the particular combination of variables on the target boundary to be equal (in an approximate sense) to the variables on the source boundary. For this technique to work well, the source boundary should correspond to the coarser mesh, and the target boundary to the finer mesh. The resulting level of accuracy will be what would be expected from the coarser mesh.

On portions of the sections that are not in contact (up to the geometrical tolerance) with a section on the corresponding boundary, a zero velocity condition will be imposed automatically, and a zero flux condition will be imposed for all other variables.

Accuracy of mass conservation is a reasonable measure of the quality of the non-conformal interface. You can check this by looking at the mass fluxes through boundaries that appear in the Ansys Polyflow listing file.