To refine the cladding a number of things needed be done to the mesh.

Simple evaluations of each triangle: measure angles and make sure no triangles were more or less than certain values.

Maximum spans of the edge profiles limited triangle sizes.

Perimeter length proved useful and was easy to get out of Rhino.

Same with triangle area; easy to get from Rhino.

The angle between triangles would affect gaskets and profiles. Ridge or Valley would also affect the node complexity.

The global position would affect deflection in joints and internal sag/bending.

Slope vector to judge water run-off which in turn affected drainage.

Node complexity; a combination of triangles/node and triangle-triangle plane angle.

Information output from Rhino, inserted into an Excel spreadsheet and sorted by area.

Summaries of triangles in Excel gave a good overview of what was going on if we did changes to the geometry.

Comparing triangle area to perimeter gave this diagram. The spikes in the light green curve indicate triangles with a large perimeter to area, ie. a triangle likely to be a long and sharp sliver.

Another way of looking at the same data.

A gaph of distribution of triangle Areas. From this we could pick out triangles that were either too large or too small. Similar graphs were done for edge lengths and Plane-Plane angles.

Via a simple lookup VBscript in Excel triangles such as the large ome on the left leg and the long, narrow ones near the folds could be located in the rhino model. Not the most elegant solution, but it worked. Wish we had more time to build some fantastic tool to do the optimization for us....