k-e model for the transition region

  • 26 Views
  • Last Post 09 April 2019
  • Topic Is Solved
LukeM posted this 09 April 2019

I am simulating a system that has flow ranging from laminar to turbulent conditions (up to a Re≈10,000). For my analysis, I have to use a k-e model for various reasons. From some preliminary testing, it seems like some of the k-e turbulence models predict turbulence setting in too early as I ramp up the Reynolds number. Is there a specific wall function or k-e model I should use to better predict this laminar-turbulent transition region? 

Order By: Standard | Newest | Votes
abenhadj posted this 09 April 2019

Do use the omega based transition models. If your device Re is already large just use full turbulent model (SST or GEKO). Stick into first node at yplus=1 And at least 15 cells across BL.

Best regards,

Amine

  • Liked by
  • LukeM
LukeM posted this 09 April 2019

Some of my UDFs use k and epsilon, which is a major reason I was sticking to the k-e models. It seems like I can still access epsilon while using the omega-based models. That also seems to be true with the SST model. Is that correct? Do I need to change anything in my UDFs to switch from k-e models to k-w/SST models? 

 

Thanks, Amine. 

abenhadj posted this 09 April 2019

Check if the variable is available to avoid any crash. If not available you need to calculate it and adjust.

Best regards,

Amine

  • Liked by
  • LukeM
Close