[rudyy]

Hello,

I have a simulation with a very large temperature distribution (between room temp and about 1500 C). The computational domain is a cylinder with some internal hardware about 70 inches long a few inches in diameter. There are some very low flows going through the cylinder (laminar regime).

Flow velocities stabilize quickly, but temperatures usually take a very long time to stabilize. With with time scale factor set to 1 usually I have to leave it overnight to get any result. Recently I got the courage to bump the TSF up to 10, then 100 and then 1000. Now my strategy is as follows:

5 - 10 iterations with TSF = 1
Bump TSF to 100 (or 1000) and iterate until temps start to stabilize
Reduce TSF to 1 and iterate until convergence

Is there any "risk" of bumping the TSF like this? It has been an effective way to converge faster (~ 200 iterations vs. 2000+ iterations) and speed up my work, and the results appear reasonable.

Thanks in advance