[ski]

Hi all,

Im using an alternating direction implicit method with individual sweeps solved directly via the TDMA to solve the P' equation (SIMPLE algorithm). I have noticed that when I sweep parallel to an outlet, on the cells adjacent to the outlet, I get some rather odd results such as flow entering the domain through the outlet and eventual divergance.

For example, for channel flow I have set a velocity inlet at the west boundary, an outlet on the east boundary, and walls at the top and bottom. If I sweep in the North south direction on the column nearest the outlet I have problems. If I do all other sweeps but miss this one out, all is fine. Can someone please explain whats going on here.


Thanks