[p1nkman]

Hello, hope this post finds everyone in good health.

I have a query regarding judging convergence. I have read through numerous number of posts in this forum however I need to double check with fluent professionals here to know if I am on the right track. Apologies beforehand as this post might sound noobish.

I have setup a solar load model (Rosseland, k-epsilon, Standard Wall Fn, Full Buoyancy Effect) with pressure inlet and pressure outlet using the pressure-based solver. I had to reduce the URF in order to get the residuals stable. The mesh skewness is below 0.88 and the Orthogonal quality is above 0.15.

Solution Method:
Least Square Cell Based, Second-Order Upwind
The URF's are:
Momentum
0.7
Turbulent Kinetic Energy
0.5
Turbulent Dissipation Rate
0.5
Turbulent Viscosity
0.7
Energy
0.7


1. The scaled residuals do not converge, staying a bit above the default convergence criteria, however I had set up a surface monitor for velocity and static temperature at a specific plane (chimney inlet) and I stopped the iteration (after about 2500 iterations) when these two variables become a straight line, i.e no longer changes.

Am I good to use these results or there are more variables that need to be ensured? My end report will be focusing on these two variables.

2. To monitor the surface integrals of Velocity and Temperate at the created plane, what report type should I use? Integral or Area-Weighted Average?

I hope my question was clear, thank you for reading.

[p1nkman]

Anyone?? Please?

[LuckyTran]

Residuals are a global measure of tolerance, how well the solution satisfies the continuity, momentum balance, etc. If they stay constant and don't meet some 10^-inf criteria that you set, you solution can still be converged but not to a tight tolerance.

Convergence should be verified using monitors and checking the values over iterations/time. If your velocity and temperature is perfectly constant, then it is obviously converged, at least for that particular quantity that you are monitoring.