Showing results for 
Search instead for 
Do you mean 
Reply

STAR-CCM+: How to test more designs!

[ Edited ]

The biggest constraint that engineers face in exploring more of the design space in less time, with respect to CAE, is the CAD to mesh process. There are multiple ways of approaching this, here are a few of them:

  1. Importing geometry
    1. The most common way is to import geometry directly from a CAD package you may already be using. If the geometry is clean (has no holes, folded surfaces, etc.) after being brought into STAR-CCM+ then setting up the simulation the first time is relatively straightforward. Some changes to the geometry (position, angle, scaling, etc) can be made within STAR-CCM+ without the need to go back to the CAD package. Some changes, however, will require that you revisit your CAD package, make the change, import the new geometry in STAR-CCM+, and swap out the old geometry in the meshing process and at the regions level. New surfaces need to be connected to the existing mesh refinements and boundaries. When the user swaps the new geometry for the old geometry, STAR-CCM+ will try to connect the new surfaces based off of their names and the corresponding naming that was used in the old geometry. So, if the new surface names are not exactly the same, the connection may not happen as you intend it to. If they are exactly the same, STAR-CCM+ should be able to make the correct connections.
    2. If you import geometry into STAR-CCM+ and the geometry has surface errors, there are a few approaches to take to get it cleaned and then swap out the part in the pipelined setup. 1) It’s possible that changing something in the original CAD program could fix this. 2) If not, users can setup a pipelined operation to surface wrap the geometry which will produce a clean geometry. This, however, takes time to setup and each wrap also takes time. It also assumes each subsequent geometry will be similar in nature so the same wrapping settings can be effectively applied. 3) The final (general) scenario would be to manually repair the surface of each new geometry that comes into STAR-CCM+. This by far would take the most time.
  2. Direct CAD to STAR-CCM+ connection
    1. CAD Clients: This directly connects STAR-CCM+ to NX, Catia, Creo, or Inventor (as of 2017). This is an additional licensing feature that allows any change made to the CAD to update in STAR-CCM+. Exposed parameters can be seen within STAR-CCM+. Requires that your CAD package has the needed libraries/add-on to communicate with STAR-CCM+. This will allow you to use Optimate/Optimate+. However, most CAD programs are Windows based and this can cause additional complexity if you are trying to run on a cluster, which are normally linux based. See below section on optimization for more information.
    2. STAR-CCM+ 3D-CAD: STAR-CCM+ has CAD built into it. No additional features are needed for this. Geometric design parameters can be exposed. An example of this is shown here: https://steve.cd-adapco.com/articles/en_US/FAQ/Class-Supplements-for-CFD-Fluids-and-Heat-Transfer-Co...

 

 

Using a direct CAD to STAR-CCM+ connection instead of importing geometry allows users to avoid spending time on swapping geometry, checking surface diagnostics to ensure the geometry is clean, and/or wrapping the geometry. Instead, users can change the CAD (enter a different number for an exposed design parameter – or move a sketch point, move a body, change a radius, length, etc within CAD), click mesh, wait for the meshing to complete, and click run. Multiple design iterations can be manually run much faster this way.

 

To further speed up this process, you can automate this design exploration process. One approach would be to develop a java macro to automatically run specific simulations (in addition to other functions like saving pictures, sim files, etc). While this does require additional licensing, it does require your time to develop this.

 

To avoid this, our optimization add-ons/software can enable you to automate simulation runs and data collection while also intelligently searching the design space using a searching algorithm. This, in short, allows engineers to find better designs, faster. It also expands the searched design space from a handful of designs, to potentially hundreds or thousands of simulations. See our optimization sections for more information.