To gain confidence in your CFD results, it's useful to run benchmarks and compare the results to data that you have from model tests, full-scale data, or even other CFD software. When doing so, it's important to be sure that you are comparing "apples to apples" in terms of the geometry of the model, the flotation condition, water properties (density and viscosity), and details such as appendages.


Here is a summary of things to consider during your benchmarking process:

  1. If you have the raw model scale test results from the towing tank (e.g., the measured model scale resistance) it is best to run the CFD simulations at model scale and compare at model scale. This helps to eliminate many of the assumptions made when expanding model scale data to full scale.
  2. If you are able to run at model scale remember to use the model scale water density and viscosity. Also, make sure that your tow point in the simulation matches that for the model test. Orca3D Marine CFD defaults to towing from the CG, but you can move the tow point if needed.
  3. The CFD grid/mesh is based off the CAD geometry from Rhino. Part of what Orca3D Marine CFD does is mesh the Rhino model before passing it to SimericsMP. If you click on the Options button in the Orca3D CFD command you will see a button to adjust the surface mesh. It’s probably a good idea to increase the density of the Rhino surface mesh to make sure you adequately capture the Rhino geometry (e.g., Initial Grid Quads to 500, Density to 1, Minimum Edge Length to 0, Max Aspect Ratio to 3). You will have to reset your face attribution after doing this. The better the quality of the Rhino mesh, the better the resulting CFD grid will be.
  4. In the same options dialog you will see the CFD Grid Size. This controls the density of the CFD grid/mesh over in SimericsMP. In general, the finer you make this setting, the more accurate your CFD answer should be, but the longer it will take to run. “Normal” is the default setting but you should consider using a finer setting for benchmarking. It's recommended that you pick one speed, say the design speed, and do a mesh convergence study where you run Normal, Med Fine, Fine (and maybe Extra Fine) and see how the resistance varies. Then you can make your decision of which setting to use with the rest of the speeds.
  5. Of course it is important that the model that you run matches what was tested in terms of things like what appendages were included and of course the geometry itself.

The mesh is likely the most important of these, though all are important. Let us know if you have any questions about any of these suggestions.