Modeling needed for illogical flows
 

I have a couple passive convection experiments which produce seemingly useful but illogical results. The concensus on sci.mech.fluids and elsewhere is that CFD would explain them but apparently no one has seen anything similar.

The first is a strongly self-confined near wall layer in otherwise turbulent conditions when there is a tuned pressure drop along the downstream edge of the heat transfer surface [http://idt.net/~ph/airwork.htm#fluwall]

The second is a self-confined convection cell between vertical planes. Smoke from a burning taper between planes 1/4" apart continually recirculates with itself rather than going to the top of the column as it does with a plate separation of 3/8" [http://idt.net/~ph/airbox.htm]

If you've seen these or similar things in CFD or the lab, or might be interested it trying to model them, please let me know.

Phil Henshaw ph@idt.net