[cfd new]

hello all. I am trying to analyse thermal stratification in horizontal pipe with bends and t-junctions.
T-junction is at the end of the pipe. In T-junction, main line is having high flow and low temperature condition and branch pipe is having very low in-flow at a higher temperature. So turbulent in-penetration is expected in the branch pipe due to high flow in the main line. Which will result in stratified layer.

I am providing information regarding my model. Mesh quality (max skewness) is 0.78. Poly hex-core mesh i have generated in fluent meshing. Min orthogonal quality is 0.2
SST k-w turbulent model
In material properties, density is given as a polynomial (function of temperature)
In operating condition, operating temperature is given as main line temperature which is low temperature. and operating density has been set as a density corresponding to the branch pipe inlet temperature which is high temperature. hence the lower density.
main line as well as branch line inlets are defined as velocity inlet and turbulent is modelled as turbulent intensity and hydraulic diameter. (branch pipe is having inlet at a distance..with very low flow..0.001 times the main line flow)
outlet is pressure outlet with default settings.

Solution parameters
method is coupled with pseudo transient
pressure descretisation is PRESTO!
all other is by default
pseudo transient time scale is set as 0.1 (Initially I have set it as 1 which is default. but error was not going down so I have decreased it)

This is a steady state problem.

ok now my question is:
error is not going down as well as physics it is capturing is also I am not sure that is it correct. I have tried the same analysis using pure hex mesh in whole domain except in T-junction. But by doing so tet generated in T-junction was coarse. so I wasn't confident that this will work. Thats why I have shifted to poly-hexcore mesh. But results are not satisfying. And I just have the instict that may be because of this mesh results is not converging.
Error just stops reducing (It is getting stable at around 0.01 in continuity) It oscillates there in small bend. and velocity near the T-junction is chaotic. It is also not getting stable.

Kindly help me with this problem.
TIA