Reverse flow starting when about to converge
I have a 3D wing which i need to analyse for lift and drag. Mach = 0.18 or V=58m/sec.
ICEM CFD determinant 3*3*3 quality of mesh is 0.8 to 1. Wing span = 0.5715m mean aero chord = 0.13m Operating conditions are ModelSA airconstant density upper, lower and front face of far fields are velocity inlet in positive x direction, with turbulence = 1% and length scale of 0.015m Back surface of far field is pressure outlet..everything is at 1atm Solution method is at SIMPLE scheme and default options Initialization done from the inlet values. The calculation proceeds smoothly until 1000 iterations. It almost starts to converge but suddenly reverse flow starts and urbulent viscosity limited to viscosity ratio of 1.000000e+05 is also displayed..they keep increasing I have no clue about the reason!!!!...Pls help.. 
what are dimensions of domain extend ? I mean X, X, Y, Y and Z, Z

X is 1.1916m to 2.6183m
Y is 1.27 to 1.27m Z is 0 to 0.991m @Far, the wing is the same one which i had posted earlier http://www.cfdonline.com/Forums/ans...required.html 
I would like to know these dimensions as chord (or any appropriate characteristic length) * X etc.

i think you wanted to know if these values are calculated based on some characteristic length??
If so, yes..the far field was created as a multiple of chord length (5) The back surface of the far field is 20times (100) from the leading edge corner. The upper lower and front faces are 10 times (50) from the leading edge corner. Just now i had changed the bunching in the vertical dir near the wing to have good Y+. I increased it by two decimal points. The reverse flow stopped but turbulent viscosity ratio problem is still there. It came after 2000 iterations..I have no idea if there is a relation between them. 
All times are GMT 4. The time now is 17:40. 