Stuck film parcel warning
 

Hi all,

I was running a Direct-injection case which has a side-mount fuel injector setup, like most common DI engines available today. The fuel injectin event started at -299 degree ATDC and continued receiving unreal large amount of stuck film warning since the begining of injection.

I have noticed that a former thread which mentioned nearly the same issues: https://www.cfd-online.com/Forums/co...imulation.html

Unfortunately no solution has been provided from this thread.

I have checked the geometry from case by diagnosis mode and everything seems fine. The solution timing is a variable timing as most Converge example case did. The fuel injection setup is identical as the available Converge GDI SAGE model.

As looking closer to issue, it seems that it would generate a stuck parcel warning if the fuel parcel hit the cylinder wall. As the dynamic mesh started to trim when piston moving up, those stuck parcel will be gone instead of vaporize or burn.

Please comment and share your thought about this issue. Thank you.

Hello Shanghua,

The "Stuck film parcel" warning messages occurs when there is some difficulty/complexity locating where a film parcel should go when they move along boundaries. The warning message is shown since if often indicates : misaligned geometry (for ex, misaligned liner), sharp corners/crevices in geometry, sliver triangles, improper seals. This is not a complete list of causes.

If you are certain, that the geometry is correct, aligned and well triangulated, you can ignore such warnings. CONVERGE does not remove the parcel when you see such warnings. You can verify this by monitoring injected fuel mass within the domain.

If you continue to face issues with your setup, please reach out to us at support@convergecfd.com. Please use your official email for all correspondence with Convergent Science. Please mention your issue, attach your case setup and add the cfd-online thread, as reference.

Sincerely,

Thank you for your reply, Kislaya.

I just double check the modeling and fixed some setup issue. Everything works fine now.

Just for curiosity, in what kind of circumstance will needed to switch 'Wall film' to 'Rebound/slide' or 'Vanish' ? I have looked through manual but having no clue about picking up these options correctly.

Hello Shanghua,

If you observe/expect liquid film to be present, then you would want to include a Wall Film model in your CONVERGE simulations. This becomes important for PFI engines, GDI injection during intake, and cold start cases, were significant amount of liquid film can be formed. Liquid films on cold walls do not evaporate as well and will affect your equivalence ratio for fuel vapor combustion. Formation of liquid films will also affect solid temperatures, creating local cold spots, which become important to evaluate for CHT simulations. Just to name a few examples. You're free to select any of the three Wall Film Splash models. O'Rourke model is typically used for IC Engines. Kunkhe and Bai Gosman are typically used for Aftertreatment/Urea-SCR simulations.

If you do not expect liquid film or feel that it should be minimal and would not impact your results much, you can use the Rebound/Slide to reduce your computational expense. For typical Diesel simulations, we do not expect wall films and the wall temperatures are typically very hot to enable film formation. You will find this selection in our Diesel example cases.

The Vanish model will just remove the parcels when they hit walls and can be helpful for some unique problems.

Hope this helps.

Sincerely,

Thank you Kislaya for your detailed reply. This is very helpful.