Hi, thanks for the link.
Some comments from my end:
- First, to address your main question, residuals are not the only thing to keep track of, when assessing convergence. Naturally they are important, but I have already seen cases that:
- Have “high” residuals but the parameters of interest have not converged and,
- Cases where the residuals are very low (e.g. <1e-6 for all parameters), but the quantities of interest have not yet converged.
If the residuals plot looks good, that’s definitely a good start, but it’s no guarantee of convergence just by itself. Sometimes you just have to let the simulation run for longer until your parameters of interest have converged.
In any case, some tips:
- For external aerodynamics cases, it’s always a good idea to lower the absolute tolerance of the turbulence solvers. The residuals from the turbulence parameters tend to drop very quickly, and we should ensure that the solvers continue to iterate:
- I’d also highly recommend manually estimating the turbulence parameters (e.g. using this calculator, there is also a very interesting discussion here).
PS: in the calculation, for external aerodynamics, the turbulence intensity is often very small (0.5%-1%). The second link contains insights about the eddy viscosity ratio as well.