Thin/double-walled geometry

Hi, I am evaluating Caedium for use in a sailboat engineering environment, One of the areas I am particularly interested in is the capability to model sails. Using the on-line examples and the Baltic 50 example in particular for guidance, I have imported sail geometry consisting of a volume with an internal surface representative of a sail. I am however not having much success in defining a double-walled (thin) surface for the sail. Is there a detailed example available showing how this is typically done?

First try a simple 2-sided case

Before trying a complicated example, I suggest you try a simple case generated within Caedium:

  1. Create an outer box
  2. Create a face fully within the box
  3. Set the Topology Only property to on for the Boolean->Subtract tool on the Geometry Tool Palette
  4. Subtract the face from the outer box - this will embed the face in the volume
  5. Apply the Faces->2 Sided tool to the embedded face to convert it into a double sided face (membrane)

For more details on topology-based operations see "Topology Operations".

Then try simulating a flow through the box over the membrane in the usual manner similar to the tutorial "Incompressible Flow Through a Pipe into a Box".


  • Importing 2 co-incident faces as a membrane is not the way to proceed. You need to explicitly convert a single face once it's embedded in a volume into a 2-sided face as outlined above.
  • The embedded face has to be in free space - no other coincident or intersecting geometry.
  • Order is important, first create the embedded face then make it 2-sided.
  • Only a face embedded in a volume can be made into a 2-sided face.

Many thanks

Many thanks for the quick reply, I have successfully implemented the suggested procedure on an imported sail surface. As a matter of interest I am using Rhinoceros to generate the sail shapes and importing the geometry resulted in the sail surface being imported as a volume. I was only able to convert the embedded sail surface once I had deleted this volume (after carrying out the Boolean subtraction).

Odd volume

Interesting. I worked on this very problem and found the same exact result, using Rhinoceros also. I am very pleased with the results from this test simulation.

I find the double sided surface 'aliases' when presenting results as a color map, the 'other side' must be suppressed somehow. I used transparency to reduce the distraction of the aliased surfaces, setting it so the iso-Cp lines were for the pressure side while the shaded color was for the suction side. Some lift monitor numbers verified a reasonable prediction for the combined total, and the Cp values made sense- they didn't before I read about deleting the volume- the first cut I did I had not deleted the strange volume created on import of the single surface via STEP. I also 'discovered' I can use the surface edges to seed the streamlines, also very useful. Would be happy to post results.

Happy to see results

For more on streamlines see "Particles or Streamlines"

"...happy to post results." check your email for contact details.


Though my volume was too small and I am certain the side walls had an effect, I use calculations to predict the velocity gradient I might expect over the water surface, the 'sail' is created based on the apparent wind along the mast to twist appropriately. Pretty good agreement, my design methodology seems to be on the right track according to the Caedium simulation!

I note in your simulation you used a cylinder for the volume, can you elaborate on the reason for that? I would have expected a huge cube...aligned with the apparent wind direction.

Cylinder Outer Boundary

I used a cylinder as the outer boundary in the example "CFD Analysis of a Sail Yacht from SketchUp", so that I could vary the wind direction using Inlet-Outlet boundary conditions. With a box that isn't aligned with the wind direction you sometimes encounter simulation stability issues caused by the corners.