Learning Forum

[dk89]

Hi all,

I have successfully validated the air flow field around a compressor rotor blade. My computational domain consists of an inlet duct, the rotor domain (hub, blade, and shroud) and an outlet duct. There are two interfaces between the inlet duct-rotor domain and rotor domain-outlet duct. I use the frozen rotor approach for the two interfaces. I inject some inert particles using the DPM. Most particles are tracked, but after some DPM iterations I get the message 'Warning: couldn't find neighbor across sliding interface 12 type 24, reflecting'. The sliding interface 12 corresponds to one side of the interface between the inlet duct and rotor domain. Moreover, I don't specify a DPM boundary condition in this interface since it is not a wall. Could you please help me with that? Thank you for your time.

[ai0013]

I think you have set up the "reflect" boundary condition for your interface if that's the case. Shouldn't it be "interior"? Try looking at the DPM bcs at your interface 12, and see if that Is enabled. If you want to specify other DPM bc's than the default ones, then you have to use UDF for that. 

[dk89]

Hi,

Thank you for your answer. I have not defined a DPM boundary condition at the interface 12 because it is not a wall. You can see the interface 12 in the attached image. The side with the ID 12 is one of the two sides of the frozen rotor interface (inlet duct-rotor blade domain). The R1-S2 interface (rotor blade domain-outlet duct) is a similar frozen rotor interface, but I didn't get any warning for this interface during the DPM tracking.

Regards

[Rob]

Are you losing many particles? 

[dk89]

Hi,

I track 350 particles totally (I know it's a small number), 348 of them escape, whereas two of them are incomplete. Thank you.

[Rob]

Then don't worry about it. It means a couple of streams have got confused at the interface. This usually means they've hit a node at an odd angle and potentially then tried to enter two cells at once.