Computational Domain Selection


#1

Dear Users,

How to select a Computational domain? I mean is there any minimum upstream length and domain diameter.

For example if we want to study stress, velocity and other pressure parameters for a ball or small aircraft .

Thank you


#2

Hi @Vinaykumar and thanks for your question!

This indeed a tricky thing to take care of and if the domain is too small the boundaries might have an influence on the flow and will distort your results giving you complete nonsense (in most of the cases). A good technique is to start with a small domain (but not too small) and see if values like drag change if you incrementally increase the domain size but keep the mesh constant assuming your mesh has a good quality with ideally 0 illegal cells to make sure distorted cells do not mess with your results.

However a ball and a “small” aircraft are different things in how you would approach such a problem. An aircraft needs a way bigger domain than a simple ball and you also have to consider if the simulation you are doing is supersonic or subsonic/transsonic.

To sum things up it really depends on your problem and experience will definitely help there to know how big your domain has to be in order to be sufficiently big to make sure boundaries do not mess up your study but also not too big because a bigger size is money i.e. core hours :wink:

Cheers,

Jousef


#3

Hi @Vinaykumar,

Just to quickly add on to what @jousefm has brilliantly mentioned. A quick and efficient way to deduce what domain size should be used for what type of external flow simulation is to just refer to literature or validation studies. They should have dictated or referenced some source on what the recommendations are for the specific type of simulation.

Hope this helps.

Cheers.

Regards,
Barry