Since you’re running a steady-state simulation:
Only the final value counts, because the simulation will converge to a single and final result (optimally speaking, of course). Looking at the values for the domain, they seem very stable:
